included in statistical reports. The popularityof these direct-market farms is increasing as aresult of the expanding local food movement.One of the premises behind this movement isthat eating locally grown food is better for theenvironment, yet little is known about howcold-climate strawberry production affectsthe environment.

The conventional matted row system(CMR) is a perennial system and has beenthe primary method for strawberry produc-tion in these regions for much of the pastcentury. More recently, cold-climate plasti-culture (CUP) (http://plasticulture.cas.PSu.edu:nraes_manual.htm) has become popular.Both CUP and CMR use methyl bromidefumigation and other pesticides to controlreplant disorders. However, methyl bromideand other traditional pesticides are being phas-ed out as a result of the restrictions mandatedin the Montreal Protocol of 1985 (http'.//www.epa.gov!ozone/intpoFindeX.htm]) andthe Food Quality Protectiots Act of 1996(http://\vw.cpa.gOv!pesticidcs!regUltttitgftipa1). prompting the development of altemativcproduction strategies to manage these pestpressures such as advanced matted row(AMR) (Black et al., 2002a. 2002b). New pro-duction practices should he evaluated forimprovements in sustainability such as theability to conserve soil and water and to reducethe movement soil, nutrients, and pesticidesfrom the production fields to nearby waterresources. From these evaluations, guidelinesand strategies can he developed and revised toaddress the standards associated with the ('leanWater Act and the implementation of the TMDL(Total Maximum Daily Load) program(USEPA, 1994, 1998).

Pesticides, nutrients, and suspended soilparticles in runoff have been shown to havesignificant negative effects on aquatic plants(Forney and Davis. 1981; tones and Winchell.1984) and estuarine organisms (Clark et al..1993: Savitz et al., 1994; Scott et al.. 1994).Bactericides, insecticides. and fungicides.which are frequently used to protect fruitcrops, are known to have more adverseeffects ott finfish, shellfish, and other aquaticorganisms at environmentally relevant con-centrations as compared with the herbicidesused on grain crops (Pait el at.. 1992). In mostcases. 1% to 6% of applied pesticides in rowcrop production systems can be transportedoff-site in runoff water and on soil particles inthe runoff (Wauchopc. 1996); however, lossesof nearly 40% of the applied pesticide havebeen observed with some vegetable produc-tion systems (Rice et al., 2002).

In general, the amount of pollutants trails-ported in runoff from fields depends largelyon the slope of the field, soil type. man-agement practices, presence or absence ofsubsurface drains, and the amount. timing,and the intensity of rainfall after applicationOf nutrients or pesticides ( Bengston et al..1990; Leonard et al.. 1979: Poling andMonks. 1994; Triplett et al.. 1978). Fieldrunoff with its pollutants enters nearbystreams affecting water quality at the water-shed level and ultimately at the basin scale.

FIOtsTSCI1\ i. 44(2):298-305. 2009.

A Comparison of Three Cold-climateStrawberry Production Systems:Environmental EffectsMatthew D. Stevens'Department (# P/ant Science and Landscape Architecture, Universitt'of Macv/and, Plant Sciences Building, College Park, MD 20742

Brent L. BlackPlants, Soils and Climate Department, Utah State University, Logan,

UT 84322-420

John D. Lea-CoxDepartment of Plant Science and Landscape Architecture. Universityof Maryland, Plant Sciences Building, College Park, MD 20742

All M. Sadeghi, Jennifer I-larman-Fetcho 2 , Emy Pfeil, Peter Downey,

Randy Rowland, and Cathleen J. Hapeman3Beltsville Agricultural Research Center, Agricultural Research Service. US.Department ofAgriculture. 10300 Baltimore Avenue, Beltst'ille, MD 20712

.4dditional index words. advanced matted row, Fragaria Xan000Ssa, erosion, pesticide, plastic

mulch, nitrogen uptake, nitrogen use efficiency

Abstract. The environmental effects of the three strawberry (Fragaria xa,zanasSa) cold-climate production systems were compared: the traditional method of conventionalmatted row (CMR) and the two more recently developed practices of advanced mattedrow (AMR) anti cold-climate plasticulture (CCP). Side-by-side field plots were instru-mented with automated flow meters and samplers to measure and collect runoff, whichwas filtered and analyzed to determine soil, pesticide, and nitrogen losses.. .Althoughannual mean runoff volumes were similar for all three production systems, the soil lossesfrom CMR plots were two to three times greater than the CCP plots throughout the studyand two to three times greater than the AMR plots only in the first year of the 3-yearstudy. In general, decreases in erosion and runoffvolumes were observed in plots thatwere disturbed less by machine operations anti had less foot traffic as a result ofdecreased need for hand weeding and in the plots that used straw mulch in the furrowsbetween the beds. Timing anti intensity of precipitation events also influenced the amountof soil erosion. Pesticide residues and nitrogen losses were also greatest in the runoff fromthe CMR plots. The two systems that used drip fertigation, AMR and CCP, also hadhigher nitrogen uptake efficiencies. Overall, the CCP and AMR systems performedsimilarly for most criteria; however, considering the nonrenewable nature of the plasticmulch and the need to dispose of the plastic mulch in a landfill, the AMR system was moreenvironmentally sustainable than the CCP system.

Strawberries presently rank fifth in theUnited States behind bananas, apples,

Received for publication 7 May 2008 Accepted forpublication 24 Sept. 2008.We thank the North American Strawberry GrowersAssociation for financial support of this research.Phillip Edmonds for field maintenance and pesti-cide application, and Dr. Dennis Timlin for assis-tance with statistical analysis.This paper is a portion of a Master's Thesis sub-nutted by M. Stevens.Use of trade names does not impl y an endorsementof the products named or criticism of similar onesnot named.'Current address: Horticulture Agent. NC ('ooper-ative Extension Service, 359 Ferrell Lane. I lalifax.Nova Scotia, Canada.Chemist, formerly with USDA-ARS.

'To whom reprint requests should be addressed:e-mail cat hleen.hapemanars.usda.goV .

oranges, and grapes in flesh fruit consunip-tion. Over 95% of the strawberries in theUnited States are cultivated as an annual cropin coastal California and in central Florida(USDA-ERS. 2007). The remainder of thereported production is from Oregon. NorthCarolina, Washington, Pennsylvania, Michi-gan, New York, Ohio. and Wisconsin withover 12.000 planted acres being reported.However, nearly every urban center in themidwest, northeast, and mid-Atlantic regionsis surrounded by small-scale fruit and vege-table farms that market produce directly tothe end consumer through roadside stands.farm markets, and customer-pick sales.Strawberries are an important early-seasoncrop for many of these small-scale opera-tions. but production data from these smallfarms represent such a small fraction of thetotal domestic production that they are not

298l-ioRTScwNcE Voi.. 44(2) APRiL 2009

CROP PRonucno\

For example .sedi merit loading i ill of lie Ches-

apeake Bay has been cited as a leading causeof submerged aquatic vegetation degradationand habitat loss for shellfish and fish species(Brush and lb gartner, 2000). Nearly 40% ofthe nutrient loadings to the Chesapeake Bayhave been attributed to agriculture (http://www.chesapeakehay.net/assess/200ohealth.htm) .Excess nutrients have led to algae bloomsand anoxic zones in the Bay and havecontributed to the decline of overall Bayhealth. Many states in the northeasternUnited States have, or are considering enact-ing. a nutrient management policy for allsectors of agriculture to reduce nutrient runoffand leaching (I-ca-Cox and Ross. 2001).

When identifying best management prac-tices for strawberr y production, soil conser-vation measures, nitrogen (N) applications,and pest management strategies need to beconsidered in a more holistic approach. Thepresent multiyear field study was carried outto compare the environmental footprint ofthree cold-climate strawberry production sys-tems: ('MR. AMP., and CC'P. Runoff, plant.and fruit samples were collected during eachof the growing seasons. Runoff samples wereanalyzed for total runoff and soil, pesticide,and nutrient losses. Plant and fruit sampleswere used to evaluate nutrient uptake effi-ciencies associated with each of the respec-tive production methods.

Materials and Methods

Site description and instrumentationThe research site was located at the Hcni'y

A. Wallace Beltsville Agricultural ResearchCenter (Beltsville, MD). The site was com-prised ofMattapex silt loam (fine-silty, mixed,active, mesic Aquic Hapludult with 1.3% to1.6% organic carbon content) and had a 5.2%

%to 7.1 north—south slope. Although thisslope range is much higher than the recom-mended slope range of 1% to 2% t'or pro-duction in the mid-Atlantic United StatesPoling and Monks, 1994; O'Dell and Wil-

liams. 2000) and 3% to 5% in tile northeastUnited States ( Pritts and Handley. 1998), thesite was purposely selected to facilitate thecollection Of runoff water. The site wasdivided into nine side-by-side research plotsotnearly identical size, measuring 13.7111 longX 6.1 111 wide, that were assigned to one ofthree replicate plots for each of the threeproduction s ystems: ('MR. AMR, and CCP.Experimental treatments were establishedusing a randomized complete block designwith blocking by location and plot history.Before this study, the plots were in a sweetcorn (Zea mai.$)—tomato (Li'copersicon escu-lentil/n Mill) rotation.

,4c/''anced matted comm. In the fall beforeYear I (Fall 2001), raised beds were formedin the AMR plots with subsurface dripirrigation lines placed at a depth of 5 to 8em; the beds were immediately seeded with awinter cover crop of hairy vetch (Uciaui//usa Roth). grain rye (Seen/c cereale),and crimson clover ( Tmi/oljiwi incarnation).Three weeks before strawberry planting,

gl yphosatc ( Roundup Original Monsanto,St. Louis. MO) was applied to the cover crop.A week later, the cover crop residue wasmowed down creating a vegetative residuelayer or mat. On 12 May 2002, dormant bare-root 'Alistar' strawberry plants were hand-planted through the residue layer. This residuelayer provided some weed suppression dur-nig strawberry establishment with the re-maining weeds controlled by hand weedingand spot applications of paradluat ((iramox-one Max; Syngcnta. Greensboro. NC'). In thefall of Year 1, the strawberry beds werenarrowed (runners removed from furrows)with directed application of paraqliat. TheAMR system is described more fully else-where (Black et al.. 2002; Stevens. 2005).

Cnni'eniiona/ matted roil'. ('NI R plotswere prepared in March of Year I . Dormantbare-root 'Allstar' strawberr y plants ( NourscFarms, South Deerfield, MA) were set at aspacing of 45 cm within row and 1.5 inbetween row centers. Overhead irrigationwas delivei'cd using gear-driven lawn sprin-klers placed along tile edges of each plot.Plots were cultivated periodically to controlweeds with hand weeding around motherplants and hand placement of runners.

Cold climate plasticim/Ium'e. In August ofYear 1(2002). CCP plots were prepared withraised beds covered with 1.25-mil blackplastic mulch with subsurface drip irrigationlines placed at a depth of 5 to 8 cm down thecenter of the beds. 'Allstar' plug plantsDavoncrest Farms, Hurlock. MD) were

planted through the plastic mulch in offsetdouble rows at a 30-cm within-row andbetween-row spacing. Weeds between bedswere controlled by directed application ofparaquat.

Fruit was harvested from each system inspring of Years 2 and 3 (2003 and 2004).After harvest in Year 2, the plots wererenovated to retain high production for asecond year. For tile CMR and A NI R systems,renovation consisted of mowing down plantsto remove leaves, application of 2.4-D (2.4-D6 Amine: NuFarni Turf & Specialty. BuirRidge. IL), and a period of dormancy fol-lowed by application of ammonium nitrate.For the C'CP systctii, renovation consisted ofremoval of leaves using a weed trimmer andremoval of excess branch crowns. Very fewplants were found to have more than fivecrowns and therefore were not considered inexcess.

Pesticide and fertilizer applicationEach production system was treated with

all bloom-ti me fungicide sprayregime following the recommendations ofthe Maryland Commercial Small Fruit Pro-duction Guide (Steiner et al.. 1999). Allfungicides were applied using art

plot sprayer equipped with a single-rowboom with three nozzles directing spray 11'omiiabove and both sides of - the row. In Year 2,azoxystrobmn (Quadri s Flowabie: Syngenta,Greensboro. NC'), benomyl (Benlate; DuPontAgricultural Products, Wilmington. DL).captan (Caplan 50-WP; Micro Flo, Lakeland,

FL), and th i ophanate-met hy I ( Topsin-NiWSB; ('erexagri. King of Pruissia, PA) wereapplied to control the following diseases:Botrytis rot (Botm'vti.s inem'ea Pers.: Fr).anthracnose (C'n//etntrielrumn aeimtatunm .1.11.Sinimonds), common leaf spot [itIi'cospliaer-e//a/'agam'iae ('Fill.) Lindau]. Phomopsis leafblight [Phomopsis oh.seio'a,i.s (Ellis & Everh.)Sutton], and leaf blotch (Gnnnionia co/nor! P.Karst. and (Jno,nonia lingo,we Kieb. ). Thesame group ot fungicides, except for benomvl,which was no longer available commercially,was used in Year 3. Specific application datesare given elsewhere (Stevens. 2005). Herbi-cides were applied to each plot as needed andin concert with other weed managementpractices appropriate for that system. A groupof insecticides was identified for potential useand plots were scouted oil regular basis.Minimal insect infestation was observed, sono insecticide applications were necessaryduring the course of the experiment.

Although between-row weed control inthe CMR system was primarily accomplishedthrough mechanical cultivation, the CCP andAMR systems were treated when necessarywith paraquat. Oil occasion, the CMRsystem also received an application of para-quiat. Additionally, the ('MR and AMP. sys-tems received one application each of 2,4-Dduring renovation alter harvest in Year 2.Paraquat anul 2.4-D were applied using abackpack sprayer and glyphosate by a boomsprayer. The rate and number of pesticideapplications is shown in Table I.

Fertilizer application rates, timing, andmethod differed in each system according torecommended practices for that system.Ammonium nitrate was broadcast iii one totsso large applications each year to the ('MR,sshereas tile AMR and CCP systems receivedweekly fertigations through the drip irriga-tion system.

Precipitation and frost protectionA weather station located within 0.5 km of

the research plots was used to measure themillie and intensity of each precipitation eventusing a tipping-bucket rain gauge. Tempera-tune, solar radiation, and wind speed anddirection were also recordeul at I 5-nun inter-vals. All runoff data reported here were fromprecipitation events with the exception of'24Apr. 2003 and 16 Apr. 2004. Runoff samplescollected on these dates were the result offrost protection activities. To protect straw-berry flower buds from t'rost that occurredbefore fruit set, overhead irrigation was ap-plied equally to all three production systemswhen the temperature dropped below 2 °C.

Runoff and groundwater collectionEarthen berms were constructed around

each plot to prevent water movement be-tween the plots and to capture runoff fromonly the three middle furrows within eachfouir-bed plot. Automated runoff samplers(Model 6700: Isco, Lincoln, NE) installedat the edge of each plot were equipped witha bubbler flow module (Model 730) andwere programmed to collect samples oil

Hom(TScir\mm Vot.. 44(2) Am'smL 2009 299

Runoff Events in 2002Flow Volume

CD cc N. N. C) cc cc 'D, cc C) C) cc cc C cc N-

03 03030303:: 030/::::cc cc cc fl cc cc N- N. cc cc cc cc cc cc

50

40

EE

300

20i-i0

ci-

10

0

Runoff events

Average Soil Load in Runoff

cc ' C) cc cc cc • cc N. cc cc cc cc N-cc N. cc ,_ cc cc N.

0303 mm 05 03030303

Runoff events

50

40

EE

300i

20 .2-000

10

0

Table 1. Pesticide applications and toxicity values.

Conventional matted row Advanced matted row ('old-climate plasticulture

LCw No. of Rate No. of Rate No. of Rate

Pesticide (mgL 'Y applications (kg aiacre) applications (kg al/acre) applications (kg acacre)

FungicidesAzoxystrobin 0.47 ppm 4 0.5 4 0.5 4 0.3

Benomyl 0.12 1 0.1 1 0.1 1 (1.1

Captan 0.056 2 0.9 2 0.9 2 0.9

Thiophanate-meihyl 24.6 2 0.3 2 0.3 2 0.3

Herbicides2,4-D 1.0 1(1(1'' I 3.9 1 3.9 - --

(ilyphosate 86 - - 1 2.3

Paraquat 32 I 0.3 4 5 4.1 3-4 2.0

'LC, ) (96 h) listed are for rainbow trout (Oncorhinchu.s ,ntki.r.$) except as noted (http: toxnet.nlm.nih.gov/index.html).

'LC,,, (96 Ii) listed are for cutthroat trout ( Onco)litnchus clot/it).aries ss dli formulation.

flow-weighted (volume) basis as runoff wasdirected through a fiberglass I-I-flume.

Subsurface samples were collected usingsuction lysimeters (Model SPS2 10031 EarthSystems Solutions. Lompoc, CA). Two lysim-eters were placed in the southern half of eachplot at a depth of 50 cm and periodic soil watersamples were collected from below, the root-ing zone using a vacuum collection device.Samples were stored at 4 °C until analysts.

Sample processing and analysisRunoff samples were filtered through a

preweighed 0.45-pm pore size membranefilter to determine the amount of total sus-pended solids front plot. The filters werebaked at 60 °C overnight and reweighed. Analiquot of each filtered runoff sample wasanalyzed for nitrate and ammonium using aflow injection analyzer (Quick Chem Auto-matic Flow Injection Ion Analyzer 48000:Laehat, Loveland, CO) equipped with acolorimetric detector. A second aliquot fromeach of the filtered runoff samples wasanalyzed for pesticide concentration usinghigh-performance liquid chromatography withmultiple reaction monitoring (MRM) dualmass spectral analysts.

The chromatographic separations wereaccomplished using a Waters (Milford, MA)Alliance 2690 quaternary pumping/automaticliquid sampler system equipped with a WatersYMC ODS-AQ (120A pore size: 53-pmparticle size) 10-em column (2 mm i.d.) heatedto 40 °C: mobile phase flow tate was 0,2mLm in. An isocratic elution with 65:25:10methanol:aqueous formic acid (0.1%): aceto-nitrile was used for azoxystrohin. henomyl,and thiophanate-methyl separations. Sampleswere also analyzed for captan, 2.4-D, andparaquat using a mobile phase gradient of50:30:15:5 to 80:0:15:5 methanol:water:aqu-cons formic acid (0.1°./o):acetonitrile (polyno-mial curve 9) over 8 rnin with a 7-mm finalhold. (ilyphosate analyses were not performedbecause glyphosate was used only once andapplied only to the AMR plots.

A Micromass (Milford, MA) QuattroUltima triple quadrupole mass spectrometersystem with positive ion atmospheric pres-sure chemical ionization was set to run inMRM mode using argon collision gas. Theionization source and desolvation zone tem-peratures were set to 120 and 500 C.

respectively. Nitrogen desolvation and conegas flows were set to 275 L/h and 100 Lb.respectively. The detector voltages ('or allanalytes except eaptan were 0.2 ,iA, 35 V.and 15 V for corona, cone, and collision.respectively. For captan. the detector vol-tages were 4 pA. 35 V, and 9 V for corona.cone, and collision, respectively. The parent

900

800

700

600

500

400

300

200

100

0

i

,J)

o 40-J

0(I)

20

0

and daughter tons (MS1 and MS2: in Da)monitored in the MRM transitions were: 2.4-D, 217.25, 171.18: azoxystrobin, 404.30,372.10; benotnyl, 192.20. 160.10: captan,302.12. 266.04: paraquat. 186.31. 171 .19: andthiophanate-methyl. 343.23, 151.14. Quantifi-cation was by least squares linear i'egressionanalysis of the peak area versus amount (ng)

Fig. I. Precipitation and average runoff volume and respective soil loss in runoff from each productionsystem per event in 2002. Asterisk after date indicates that difference is significant between at least twotreatments in runoff eent. (MR conventional matted row: AMR ' advanced matted row: (CP =Cold-Climate plasticul tore.

FlosI SCIENCE VOL. 44(2) APRIL 2009

iniccied on-column usliin tour extcmal stan-clard calibration curves. The limits of detectionwere 0. I pph for azoxystrobin. henomyl. andthiophanate-methyl: 12 pph for captan andparaquat: and 170 pph for 2.4-D.

Plant carbon and nitrogen analysisAfter fruit harvest in Years 2 and 3, the

plants within four 0.093 m 2 (1 sq ft) sectionsof each plot were destructively harvested.These squares were used as representativesample areas as a result of the differences inplant spacing and growth habit among sys-tems. For the CCP system, one mother plantwas present in each harvested square,whereas the CMR and AMR squares con-tained one mother plant along with severaldaughter plants. Plant tissue from each har-vested square was then separated into lea(crown, runner, root, and fruit truss. Freshweights were taken for each tissue type andthe number of leaves, crowns, trusses, andrunners was recorded. Runners arid trusseswere liirther separated to isolate leaves asso-ciated with each, thus designated as runnerleaves and truss leaves with all other leavesdesignated as crown leaves. Leaf area wasthen taken oil leaves, truss leaves, andcrown leaves. A 10-leaf subsample was takenfrom all crown leaf samples (with the excep-tion of one sample, which only had 13 totalleaves), and a subsample was taken for anyother tissue sample exceeding 40 g freshweight.

Tissue subsamples were frozen in liquidnitrogen, stored at —80 °C. and then freeze-dried using a LabConco Model 77550 Lyoph-ilizer (LabC'onco Corporation. Kansas City,MO). The remaining plant samples wereoven-dried at 60 °C for 72 Il. Dry weightsof all samples and subsamples were recordedafter drying. Freeze-dried subsamples werethen milled with a high-speed mill equippedwith a I .0-nlm screen (Tecator ('yclotec 1093sample mill: Rose Scientific Ltd., Edmonton,Alberta. Canada) and analyzed for totalcarbon and N content using an elementalcombustion system ECS Model 4010 (Cost-ech Analytical Technologies Inc.. Valencia.CA) that was calibrated using a peach leafstandard (Standard Reference Material 1547:National Institute of Standards and Technol-ogy. (jaithersburg. MD).

Statistical analysisData for plant tissue N and runoff losses of

soil. N, anti pesticides from all three pro-duction systems (treatments) were analyzedas a randomized complete block design usingthe Proc Mixed routine of the SAS programpackage (version 9: SAS Institute Inc., 2006).Treatment means were separated using thePDIFF option of the LSMEANS statement.

Results

Pi'ecipitation and runoff Ruiloff volumeswere measured per plot, and for each event,the average runoff volumes are shown inFigures I through 3. The duration, intensity,and total volume from each precipitation

evcni were also calculated from weatherstation data. Runoff from all 42 precipitationevents was measured and anal yzed during tile3-year experiment, with 16, 18, and 8 runoffevents occurring in 2002. 2003. and 2004,respccti ely. Two frost protection eventsoccurring oil Apr. 2003 and 16 Apr.2004 were also measured. Events occurringbefore 29 Aug. 2002 do not include the ('Ci'plots because these were not established uilti!mid-Aug. 2002.

In 2002, precipitation ranged from I to45 mm per event, tile ascrage precipitationamount was 20 nun, and the median was 14mm (Fig. I). The average runoff volumes perevent ranged from 4 to 875 L, 0 to 659 L, and38 to 741 L for the CMR, AMR. and ('CPplots, respectively. Although large variationswere observed, the average runoff volumeswithin each event were not significantly dif-ferent. The 2002 annual mean (and medianrunoff volumes for the CMR, AMR. and CCPplots were 214 L (136 L). 213 L (145 L). and352 L (261 L), respectively. The annual meanrunoff volumes for tile CMR and AMR plotswere not significantly thllerent: however, theannual mean runoff volumes ('or both theCMR and AMR plots were significantly lessthail the annual nleail runoff volume for CCP(P -- 0.01) (Table 2).

Three precipitation events in 2003 (Fig. 2)were of exceptional size, duration, and/orintensity and resulted u excessively highaverage runoff vollulles (greater than1000 L) and high soil erosioil (greater tilail75 kg . ha ): 22 .July (34.5 mm over 3 h), IIAug. (19.6 mm over 0.5 h). and 18 Sept. Thelast of these was the result of Tropical StormIsabel (29 mm over 13 Ii). Although theprecipitation intensity during tile tropicalstorm was mucil less than the sudden downPours on 22 July and II Aug., the large

volume of runoff was the result at' theduration of the event, high wind speeds, andsaturated higher antecedent soil moisture as aresult of two earlier precipitation eventsduring that week. Thus, tile following obser-vations and calculations exclude TropicalStorm Isabel. Precipitation in 2003 rangedfrom to 35 mm per event, the averageprecipitation amount was 17 mm, and themedian was 18 mm. The average ruiloffvolumes per event ranged from 0 to 3560 L.0 to 2960 L, and Ii to 2290 L for the ('MR.AMR. and CCP plots, respectively. Signifi-cant differences (P 0.05) were observed inaverage runotl volumes for two events (Table2): average volumes from the AMR andCMR plots were greater than the CCP plotsoil Aug. and the average runoff volumefrom the AMR plots was greater than theCCP plots on 27 Aug. The 2003 annual mean(and illedian) runoff volumes for the CMR.AMR, and ('(P plots were 551 C (161 L),539 L (358 L), and 42! I. (263 1.), respec-tively. The annual mean runoff volumes forthe CMR. AMR. arid ('CP plots were notsignificantly different.

Runoff events for 2004 (Fig. 3) werecollected only during the spring and endedwhen harvest was completed in mid-June.Thus, the measured precipitation antI runoffwere much less as the larger mid to latesimmer storms were not included. The aver-age precipitation forfor tile events measured in2004 was 13 mm and the median was 12 mm:the events ranged from 6 to 23 mm. Theaverage runoif volumes per event rangedfrom 4 to 408 C. 0 to 509 C. and ii to 476L for the CMR, AMR, and CCP plots,respectively. Average runoff voluilles perevent were significantly different for 'ourevents in 2004: 25 Apr., 7 May, 18 May.and II June. On 25 Apr., the average runoff

Table 2. Runoffevents with significant differences associated with variables measured between productionsystems (treatments).'

Runoff event(date) Variable Treatment I Treatment 2 P29 Apr. 2002 Soil loss ('MR AMR13 May 2002 Soil loss (MR .AMR20 May 21)1)2 Soil loss CMR AMR5 Aug. 2002 Soil loss CMR AMR6 Aug. 2002 Sail loss CMR AMR9 May 2003 Azoxvstrohin residues AMR CCP

Azoxy sirohin residues (\l R ('CP22 July 2003 Azoxysti'ohin residues ('MR CC PII Aug. 2003 Runoff volume AM R ('(P **

Soil loss AMR ('(PSoil loss ('MR ('(PA,oxvstrohin residues C \1 R C'01

27 Aug. 2003 Runoff volume ('MRRunoff volume ('MR ('Cl'Soil loss AMR ('Cl'

25 Apr. 2004 Runoff volume CMR AMR7 May 2004 Runofivolurne CMR CCP

Runoff volume AMR C'CP18 May 2004 Runoff volume AMR CC'P

Runoff volume ('MR CCP **II June 2004 Runoff volume AMR CC'P

Runoff volume ('MR CCP.Azoxystrohin residues AMR C'PAzoxvstrobin residues ('MR CCP

'Iligher treatments are bolded: P< 005 unless noted (**P <0.01 ***/) <.0.00!).

HORTSCIt/NCE VOL 44(2) Anne 2009 301

3000

2500

2000

1500

1000

500

C

LL

CC0-J

0tl)

90

75

60

45

30

15

Runoff Events in 2003Flow Volume

.O N ) 0 N- 01 N N O N • N flCi £3 CiCi £23503 £33503 £3 £3 :N N O N' N- 35 35 N N N

Runoff events

N U) N- N CO CO N CU N CC) ' ' N CC) U)

£3 Ci £30303 £303 ci ciN CU U) U) CO CO N- N' 35 35 0) 0) 0)

Runoff events

volume from the AMR plots was significantlyless than the average runoff volume from the('MR plots; and on the latter three dates, theaverage runoff volumes fi'om both the AMRand ('MR plots were significantly less thanthe runoff volume from the CCP plots (P <0.05; see Table 2). The 2004 annual mean(and median) runoff volumes for the CMR,AMR. and CCP plots were 128 L (38 L), 116L (13 L). and 178 L (148 L). respectively.Again, the annual mean runoff volumes forthe CMR. AMR. and CCP plots were notsignificantly different.

Soil erosion. Soil loss (kg . ha ') was cal-culated based oil total volume of runoffwater collected per plot per runoff event. thesize of each plot, and the mass of lilterablesediments per volume of runoff. Average soillosses in 2002 per event ranged from 0.1 to270 kgha . 0.1 to 21 kgha . and 1.3 to 58kg-ha I for the CMR, AMR, and CCP plots,respectively (Fig. I). The average soil loss forthe CMR plots was significantly greater thanthe average soil loss for the AMR plots in fiveindividual events in 2002: 29 Apr., 13 May,20 Ma y , 5 Aug.. and 6 Aug. (Table 2). All

40

EE

30;0S

20(.1S0

10

3

43

EE

30;0S

20 .-0a,0

10

0

these events occurred before the establish-ment of the CCP plots in mid-August. Theannual mean (and median) soil losses for theCMR. AMR, and CCP plots were 39 kgha'(15 kg'ha '). 6 kg'ha' (4 kgha '), and 23kg'ha ' (7 kg-ha '). respectively. For 2002,the annual mean soil loss for the CMR plotswas significantly greater than the annualmean soil loss for the AMR plots (P0.05); however, neither the ('MR nor theAMR annual mean soil loss was significantlydifferent from the CCP annual mean soil lossfor the CCP.

In 2003, the average soil losses per eventranged from 0.4 to 180 kg-ha - ', 1.1 to 20))kg'ha'. and 0.2 to 30 kg . ha ' for the CMR,AMR, and CCP plots, respectively (Fig. 2).The average soil loss for the CC'P plots wassignificantly less than the average soil loss forthe AMR plots in two individual events (11Aug. and 27 Aug.) and in one event for theCMR plots (11 Aug.) (Table 2). The annualmean (and median) soil losses for the CMR,AMR. and CCP plots were 27 kg-ha(4 kg-ha ). 30 kg ha ' (7 kg'ha '). and 8kg'h;t (5 kg-ha ). respectivel y . In 2003. the

annual mean soil losses for both the CMR andAMR plots were significantl y - reater than theannual mean soil loss for CCP (P < 0.05):however, the annual mean soil loss for the('MR plots was not significantly different fromthe annual mean soil loss for the AMR plots.

Average soil losses per event in 2004ranged from 0.03 to 5.4 kg-ha ', t).Ol to8.1 kg.ha . and 0.1 to 4.4 kg-ha ' for theCMR. AMR. and CCP plots, respectively(Fig. 3). The average soil losses within eachevent were not significantly different. Theannual mean (and median) soil losses for the('MR. AMR. and CCP plots were 1.9 kgha'(0.9 kg . ha '). 2.0 kg-ha '(0.4 kg'ha '). and1.9 k9-ha '(1.5 kgha'). respectively. Theannual mean soil losses for the CMR, AMR,and CCP plots were not significantly different.

Pesticide juno/f; No insecticides wereused over the course of the study becauseinsect pressures were minimal as was deter-mined durin g regular scouting inspections ofthe plots. Four fungicides, captan, benomyl,th iophanate-methyl. and azoxystrobin, wereused to control disease (Table I ). Azoxystro-bin was applied twice in both 2003 and 2004(Fig. 4). The ranges of the average loads inruttoff per event, where azoxystrobin con-centrations were above the limit of defection.were 0.1 to 384 m ­ -ha '. 0.1 to 116 mghaand 0.1 to 92 mg'ha 'for the CMR. AMR.and ('CP plots, respectively. The differencesin loads were significant t'or four events(Table 2). For the events of 9 May 2003and II June 2004, the average azoxyslrohinloads from the CCP plots were greater thanthe average loads from both the CMR and.AMR plots: and for the events of' 22 July2003 and II Aug. 2003, the loads weregreater from the CMR plots than the loadsfrom the AMR and CCP plots, respectively.

l3enoinyl was applied only once in 2003and was detected in most of the runoff eventsin 2003 after application (Fig. 5). The rangesof the average loads in runoff per event,where benomyl concentrations were abovethe limit of detection, were 3.08 to 532mgha', 2.46 to 156 mg . ha ', and 2.46 to299 mg-ha' for the CMR, AMR. and CCPplots, respectively. Captan was used twicebut was not detected in any runoff samples atany time. Th iophanatc-mcthyl was appliedonce in 2003 and 2004 (Fig. 6). The ranges ofthe average loads in runoff per event, wherethiophanate-methy I concentrations were abovethe limit of detection. were 0.1 to 464 mg-ha0.1 to 152 in-ha', and 0.4 to 329 mgha' forthe CMR. AMR, and CCP plots, respectively.Neither the average benonlyl loads per eventnor the average thiophanatc-mcthyl loads perevent were significantly different among theCMR, AMR, and CCP treatments.

Three herbicides were used during thestudy : 2.4-D. paraquat, and glyphosate(Table 1). Cilyphosate was used only on theAMR plots and was applied before activationof runoff samplers: therefore, no data existfor glyphosate loss by runoff. 2.4-D wasapplied once to the CMR and AMR plotsduring renovation but was not detected in anyrunoff samples. Paraquat was applied once to

Fig. 2. Precipitation and average runoff volume and i-espect l ye soil loss in runoff from each productionsystem perevent in 2003, Asterisk after date indicates that ditidi'encc is significant between at least twotreatments in runoff event. CMR - conventional matted row: AMR = advanced matted row: CCP =cold-climate plasticul ttlre.

302 I loin Sctmici: VOL. 44(2) Asuui. 2009

50

40

EE

30 c0

Ca

20 0-

C,)a)I-0

10

0

Runoff Events in 2004Flow Volume

U)cc CO

to 10

Runoff events

50

-c0)

"Iin0

0

2

0Cs (0 5) iN N- CC in

cc cc in Ce

Runoff events

Fig. 3. Precipitation and average runoff volume and respective soil loss in ruiioft from each productionsystem per event in 2004. Asterisk afterdate indicates that ditThrence is significant between at least twotreatments in runoff event. ('MR = conventional matted ross: AMR = advanced matted 'ow: CCP-cold-cl iinatc plasticulture.

Azoxystrobin Load per Runoff Event

100.001

E. 80.0

060.0

40.0

20.0

0.0C') C') C') ' (N (N (N (N i") (N cc cc cc cc cc cc

>. Ce N- N- 75 ci C) C , - >' >' >' ci

°' <5 (5 iS <5 CS,

C C — ' CO CS ni:

cccc <thr'5m(N(N Cs

C'S

Date of Runoff Event

Fig. 4.,Ase age atoxvstrobin load in runoff from each production s ystem per precipitation es Cot.Azoxvstrohin was applied oil Ma y 2003. 16 May 2003. 12 Apr. 2004. and 26 Apr. 2004. Asterisk afterdate indicates that difference is si gnificant between at least two treatments in runoff event. ('MRconventional matted row: AMR advanced matted row: ('(P cold-climate plasticulttire.

the (' vl R plots and multiple ti riles to hot litheAMR and CCP plots but was only detected inone runoff e cm. No significant differences inpat'aquat losses svet'e observed between p'°-duction systems for this event (data not shown).

Ait;'ogen runoff and /1/ant uptake. Thefirtilizer method, timing, and rate of appli-cation were different for each productionsystem: thus, the results for ammonia andnitrate/nitrite N runoff from individual eventscannot be compared directly. The role ofnutrients in the overall environmental foot-print of each of the three practices wasconsidered by examining the total N applica-tions. N losses, plant uptake of N. and plant Nuptake efficiency, which are shown inin Table3. Attempts to measure leached N in the soilprofile were not successful because raitithlland irrigation were not sufficient to causeenough percolation to the suction lysi metersor adequate sample volume. Some amount

of N in each of the three production systemswas taken up by the fruit. Black et al. (2005)reported 0.666 mg N/g fi'uit (fresh weight) for'Allstar' in the AMR system. Fruit N uptakewas net accounted for in this study as a resultof the timing of destructive tissue harvest.Furthermore, a complete mass balance of Nwas not possible in all practices, unrecoveredN was presumably lost either through arnrrto-nia volatilization and/or through denitrifica-tion. although neither process was evaluatedor measured.

The CMR plots received several largebroadcast applications of NI 1 4N0 annu-ally, whereas the CC[' and AMR plotsre-ceived small, weekly applications of NH4NO1through a drip irrigation system. Althoughthe N applicattoti rate was similar for all threeproduction systems, the total amount of Napplied per plant was greatest in the ('MRplots followed by the AMR plots and then the('('P plots. Sigtiificantl y higher losses of' Nwere observed in runoff from the ('MR plotsas compared with N losses in runoff li'otiboth the AMR and ('CF plots: N losses inrunoff from the AMR plots were also signif-icatitl y higher than the N losses fi'om the ('('Pplots. Although the total N applied per plantwas lowest in the CCP plots, plant uptakeefficiency was signifIcantly higher than theother production systems: ('MR plots had theleast efficient plant N uptake.

Discussion

The effects of three strawberry productionsystems on environmental resources wereexamined to this study. In general. the differ-ences among the CMR, AMR. and CCPsystems were found to be statistically moresignificant with smaller rutioff'events primar-ily because of the low variability betweenreplicates. Likewise, for larger events, thedifferences were less significant presumablyas a result of the smaller variation among thethree practices relative to other factors. Theresults suggest that the quantity and quality(intensity and duration) of the precipitationevent and timin g of the precipitation relativeto specific agronomic events (for example.

600

500

400a)

300

0Li 200

100

0

10

8 40

EE

300

0

HORTSCirNCT Voi, 44(2) A p srt. 2009 303

0)E

0-J

E0Cdi

ED

550

500

450

400

350

300

250

200

150

100

50

Benomyl Load per Runoff Event

CO CO CO CO CO CO CO CO CO C')C) C) Ci C) Ci C) C) C) C) C)

< 2 2 7 7m ' o 00Cq,CS (N ,CN

Date of Runoff Event

Fig. 5. Average henomyl load in runoff from each production system ice per precipitation c ciii. Benomylwas applied oil Apr. 2003. CMR = conventional matted row: AMR = advanced matted row: CCPcold-climate plasticult Lire.

Thiophanate Methyl Load per Runoff Event

C') CO CO CO CO CO CO C') CO CO 'IT It

C) 9 C) C) C) C) C) 17 C) C:1 C

C) C) Ci C C) C)

(5 iS (5 ( C. C.03cDcaD7 7 < < Cl) < 7 7

d i (N C5 r- CS R o,(N ,- (N

Date of Runoff Event

0)E

200

0-J

150

di

£1 100C

CL 500

I-

tilling, weeding, and formation of the beds)affected the soil and pesticide losses andrunoff volume more than the type of pro-duction system.

This phenomenon was most effectivelydemonstrated by the events of Summer 2003.

Before the 22 July 2003 precipitation event,the CMR plots were cultivated and before theII Aug. 2003 event, the AMR plots werecultivated. Concomitantly, average soil los-ses were lower from the CCP plots, whichwere undisturbed. In April and May 2002,

significantly greater soil losses were ob-served from the bare soil CMR plots relativeto the AMP. plots, which were covered by acover crop residue layer. Oil Aug. 2002. theCMR plots lost 10 times more soil than theAMP. plots, and again, the CMR plots lostsignilicantly more soil oil Aug. 2002.because cultivation was performed beforethese events. The CCP plots were installedoil Aug. 2002 and large soil losses fromthe CCP were observed in most of the sub-sequent runoff events in 2002. In all theseevents, larger sod losses were observed fromthe plots that had been most recently tilled.

The environmental footprint of produc-tion is also ti function of the pollutants re-leased to the surrounding ecosystems. Thedischarge of nutrients from production Oddsand into nearby streams and larger waterbodies is a major concern in areas wherecold-clim:ttc strawberries are produced. Tim-ing and placement of fertilizer N were foundto be critical in the CMR plots, whichhad lower N uptake efficiencies and higheramounts of N in the runoff iv. Alternately,both the AMR and CCP plots used dripfertigation, which is buried several centime-ters below the soil surface and had signifi-cantly higher N uptake efficiencies and lowerN in the runoff compared with the ('MR plotswhere the fertilizer was broadcast.

[it to the nutrients and sedimentsin the runoff', pesticides call the runoffmore toxic. The toxicity of the runoff isdependent on the pesticide concentration,the species affected, and the node of actionof the pesticide. In general. herbicides areless toxic to most nonplant aquatic organismsas compared with fungicides (Pait et al.,1992). The fungicides used in this study havea moderate to high acute toxicity for aquaticorganisms with LCa 1 >s (96 It) ranging from 50pph to 25 ppm for rainbow trout (Table I ). Incontrast, the LC 50s for the herbicides rangedfrom I to 100 ppm. The concentrations of thepesticides measured in the runoff in this studywere generally well below these ranges.Furthermore, runoff samples with maximumconcentrations of the fungicides (0,0252ppm, 0.0515 ppm. and 0.3305 ppm forazoxystrobin, bcnomyl, and thiophanate-methyl. respectively) were at least allof magnitude lower than the reported LC1s.Pesticide losses in this experiment weregenerally limited from the AMR plots ascompared with the CMR and C'CP plots: thegreatest pesticide losses occurred mostlyfrom the CMR plots.

This study does not include the envi-ronmental issues of the manufacturing ofpurchased inputs such as plastic mulch. driptape, fertilizers, or pesticides nor does it in-clude the disposal of the plastic mulch anddrip tape. The results indicate I I that the(i'MR system was least effective in control-ling soil and pesticide losses; 2) that overall.the AMR system was most effective forerosion control in the first year, whereas theCCP system was generally better in thesecond year: and 3) that the lowest pesticideresidues occurred in runoff from the AMR

Fig. 6. verage ihiopliaitate-methyl load in runolf from each production s y stem per precipitation event.Thiophanate-methyl was applied on 13 May 2003 and 19 Apr. 2004. CMR = conventional matted row:AMR = advanced matted row: CCP = cold-climate plasticulture.

Table 3. Mass balance of nitrogen (N) components: N application mate. N applied per plant, total plant Nuptake, total N in runoff total N recovered, and plant N uptake efficiency fbr three strawberryproduction systems.'

Application Total N Total N Total plant Plant N uptake Total Nrate applied runoff N uptake efficiency recovered

Treatment (kgha N) (g/plant) (g plant) (giplant) (N) (%)CMR 170.4 10.77 0.22 a 1.73 16.1 c 18.1 cAMP. 229.8 6.49 0.11 b 2.01 31.0 b 32.7 hCCP 173.7 2.33 0.01 e 1.78 76.4 a 76.8 it

'Means followed by different letters are significant (P < 0.05).CMR conventional matted row; AMR = advanced matted row• , CCP = cold-climate plastmeulture.

Fit usrSci i•N( 1 VOL. 44(2) A situ 2009

hoh-

systetli. I hese ohser\ ations suggest that theAMR and CCP systems have less negativeeflècts on our natural resources than theCMR system. Furthermore, given that theplastic mulch is nonrenewable and must bedisposed of in a landfill, the AMR system is amore environmentall y sustainable productionsyste iii.

Literature Cited

Ben gston, R . t.., I . VI. Southwick, G. H. Willis. and(h. Carter. 1990. The influence Of subsurfacedrainage practices oil losses. Traits-actions olihe Anierican Society ofAgriculturalEngineers. 33:41c 418.

Black, B.L.. J.E. [ntis. and S.C. Ilokanson. 2002a.Advancing the matted-row strawberry produc-tion system, p. 112-115. In: I lokanson, S.C.

and A.R. Jamieson (eds.). Strawberr y researchto 2001. ASHS Press, Alexandria, VA.

Black, I3.L.. J.M. Elms, and S.C. Hokanson. 2002h.A comparison of temperate-climate strawberryproduction systems using eastern genotypes.IlortTechnology 12:670 675.

Black. DL., S.C. I lokanson. and K.S. Lewers. 2005.Fruit nitrogen content of sixteen strawberrygenotypes grown lit advanced matted rowproduction system. I lortScience 40:1190- 1193.

Brush. G.S. and W.B. Hilgartner. 2000. Palcoccol-ogy of submersed macrophytes lit upperChesapeake Ray. Ecol. Monogr. 70:645 667.

Clark. .......MA. Lewis. and A.S. Pait. 1993.Pesticide inputs and risks in coastal wetlands.Env iron. Toxicol. ('Item. 12:2225 2233.

Forney . D.R. and D.E. Davis. 1981. Effects of lowconcentrations of herbicides oilaquatic plants. Weed Sci. 29:677 685.

Jones, T.W. and L. Winehell. 1984. Uptake andphotosynthetic inhibition by atrazine and its

deiLradatiolt products on Oiur species O f sub-merged vascular plants. .....nvuon. Qutil.13:243 247.

Lea-Cox. J.D. and D.S. ROSS. 2001. A review of thefederal clean water act and the Maryland waterquality improvement act: The rationale fordeveloping a water and nutrient managementplanning process for container nurser y andgreenhouse operations. Journal ofEnvironmen-tal Horticulture. 19:226 229.

Leonard, R.A., (lW, Langdale, and W.G. Fleming.1979. Herbicide runoff from Upland Piedmontwatersheds Data and implications for model-ing pesticide transport. .J. Environ. Qual.8223 229

ODeIl. C.R. and Williams. J. 2000. Hill systemplastic mulched sti-ass berry production guidefor colder areas. Publication 439-018, VirginiaCooperative Extension, Virginia Tech, Blacks-burg, VA.

Pail, AS., A.F.D. Souza. and D.R.G. [arrow. 1992.Agricultural pesticide use in coastal areas: Anational sunimaiy. 112 pp. NO.AA/NOS/ORCA, Silver Spritig, MD.

Poling. F.B. and I).W. Monks. 1994. Stoiwbertypltistic Lill Lire guide inr North Carolina. NCCES.AG-SOS. Sian. 2009. 'ltttp: www.smallfniits.orgStrawbenies production ;StPIastG.ncplas.titni>.

Pritts. M.P. and D.M. Handley . 1998. Strawberryproduction guide for the northeast, midwest,and eastern Canada. NR.AES 88:1 62.

Rice, P.J.. L.L. McConell, L.P. Heighton. A.M.Sadeghi, A.R. Isensee, A.A. Abdul-Baki, I.A.Harman-Fetclto. and Cl. l-lapeman. 2002.Transport of copper in runoff from fresh-market vegetable production using polyetliyl-enc mulch or a vegetative mulch. En viron.Toxicol. Chem. 21:24-30.

SAS Institute Inc. 2006. SAS onlitie doe. Version 9.SAS Institute Inc. C'aty, NC.

Suviz. .1.1).. D.A. Vs right. atid R.A. Smucker. 1994.Toxic effects of the insecticide diflubenzuron(Di nit liii) oil va I and development ofNapul ii of the Estuarine Copepod. Eiiivenioraa/tints. Mar. t:nviron. Res. 37:297 312.

Scott, Cl.. M.l I. Fulton, M.C. Crosby, P.B. Key.J.W. Daugornah. iT. Waldren, F.D. Strozier.U.J. Louden, G.T. Chandler, T.F. Bidleman,K. L. Jackson. T.W. Hampton. T. Huffnian, A.Shulz. and M. Bradfbrd. 1994 Agriculturalinsecticide runoff effects on estuarine organ-isms: (on-elating laboratory and field toxicitytests. ecopltvsiologv bioassays, and ecotoxico-logical hionionitorin g . EPA 600/R-94/004:288.

Steiner, P.W. R. I letlehowcr. and R. Rouse. 1999.Maryland commercial small fruit productionguide. MCES 242, 8 pp. University of Mar y

-land Cooperative Extension, College Park,MD.

Stevens. M.D. 2005. Sustainahilitv of cold-climatestrawberry production s y stems. MS Thesis,University of Maryland.

Triplett G.B.Jr., B.J. Conner, and W.M. Edwards.1978. Transport of atrazine and sin,atine inrunoff from conventional and no-tillage corn. I.Ens iron. Qual. 7:77-84.

USDA-FRS. 2007. U.S. strawberry harvested acre-age, y ield per acre, and production. 13 States.1970 2006. 5 .)ail. 2009. <http:/:usda.mannlib.cornell,cdu/usda!ers/95003/tahle04.xls>.

USEPA. 1994. National Water Quality Inventory:1994 Report to Congress. 5 ian. 2009. <hiitp:Ywww.epa.gov . 3055 94reportindex.html:>.

USEI'A. 1998. Water Quality Assessment andTotal VI ax i mum Dail y Loads Information(All AINS).5 Jan. 2009. <http:.'www.epa.gov.\vaters/ir/>.

Wiiuichope. R. 1996. p

esticides in runoff: Measure-rnent. modeling. and mitigation. .1. Ens i roil.Sci. I lealth 1331:337-344.

HORFSLIE\d- y VOL. 44(2) Aiim. 2009

305