Content uploaded by Denis Thiery
Author content
All content in this area was uploaded by Denis Thiery on Dec 01, 2017
Content may be subject to copyright.
J Appl Ecol. 2017;1–10. wileyonlinelibrary.com/journal/jpe
|
1
© 2017 The Authors. Journal of Applied Ecology
© 2017 British Ecological Society
Received:20April2017
|
Accepted:23October2017
DOI:10.1111/1365-2664.13034
RESEARCH ARTICLE
Deployment of organic farming at a landscape scale maintains
low pest infestation and high crop productivity levels in
vineyards
Lucile Muneret | Denis Thiéry | Benjamin Joubard | Adrien Rusch
1INRAUMR1065SantéetAgroécologie
duVignoble,ISVV,UniversitédeBordeaux,
Bordeaux-Sciences-Agro,Villenaved’Ornon
Cedex,France
Correspondence
AdrienRusch
Email:adrien.rusch@inra.fr
Funding information
RegionAquitaine;AgenceFrançaisepourla
Biodiversité;Ecophyto&theFrenchNational
FoundationforResearchonBiodiversity
HandlingEditor:AilsaMcKenzie
Abstract
1. Organicfarmingisapromisingwaytoreducepesticideusebutincreasingthearea
underorganicfarmingatthelandscapescalecouldincreasepestinfestationsand
reducecropproductivity.Examiningtheeffectsoforganicfarmingatmultiplespa-
tialscalesandindifferentlandscapecontexts onpestcommunitiesandcroppro-
ductivityisamajorstepintheecologicalintensificationofagriculturalsystems.
2. Wequantifiedtheinfestationlevelsoftwopathogensandfivearthropodpests,the
intensityofpesticideuseandcropproductivityin42vineyards.Usingamulti-scale
hierarchicaldesign,weunravelled the relative effects of organicfarmingatboth
field and landscape scales from the effects of semi-natural habitats in the
landscape.
3. Atthefieldscale, pest communities did notdifferbetweenorganicandconven-
tionalfarming systems. Atthelandscape scale,increasingthearea underorganic
farmingdidnotincreasepestinfestationlevels.
4. Threeoutofsevenpesttaxawereaffectedbothbylocalfarmingsystemsandthe
proportionofsemi-naturalhabitatsinthelandscape.Ourfindingsrevealedthatthe
proportionofsemi-naturalhabitats reducedpestinfestationfortwooutofseven
pesttaxa.
5. Organicvineyardshadmuchlowertreatmentintensities,verysimilarlevelsofpest
controlandequalcropproductivitylevels.
6. Synthesis and Applications.Our resultsclearlyindicatethatpolicies promotingthe
developmentoforganicfarminginconventionalvineyardlandscapeswillnotlead
togreaterpestanddiseaseinfestationsbutwillreducethepesticidetreatmentin-
tensityandmaintaincropproductivity.Moreover,theinteractionsbetweensemi-
natural habitats in landscape and local farming practices suggest that the
deploymentoforganicfarmingshouldbeadaptedtolandscapecontexts.
KEYWORDS
biologicalpestcontrol,conventionalfarming,landscapecomplexity,organicfarming,pathogens,
pestcommunity,pesticideuse,TreatmentFrequencyIndex,vineyard,yield
2
|
Journal of Applied Ecology
MUNERET ET al.
1 | INTRODUCTION
Theintensificationofagriculturethatstarted60yearsagoinindustri-
alized countries has had severalnegative impacts that limit the sus-
tainabilityoffoodproductionsystems.Althoughithasbeensuccessful
inmeetingthegrowingdemandforfood,suchanintensificationjeop-
ardizes the environment and long-term production goals, as well as
humanhealth(Foleyetal.,2011).Profoundmodificationstoagricultural
systems,suchasreducingagrochemicaldependency,whilemaintaining
cropproductivity(Tscharntke,Clough,etal.,2012),arethusneeded.
Managinghabitatdiversityatdifferentspatio-temporalscalesisa
promisingwaytolimitpestpressureandreducepesticideuseinagro-
ecosystems(Letourneauetal.,2011).Diversificationschemesreduce
pestpopulationsthrough(1)thedirectbottom-upeffectsofresource
diversificationonpest populations,mediatedbyphysicalorchemical
confusion, that limit plant host localization; or (2) the indirecttop-
downeffectsofdiversificationonpests,mediatedbynaturalenemies
thatbenefit fromalternativehostsorprey,pollen,nectar,refugesor
micro-habitatsinmorediverseenvironments(Letourneauetal.,2011).
Takingmultiplescalesintoaccount,fromthefieldtothelandscape,is
ofmajor importance tounderstandpestpopulation dynamics, natu-
ralenemyactivityandthelevelofbiologicalcontrol(Chaplin-Kramer,
O’Rourke,Blitzer,&Kremen,2011;Rusch,Valantin-Morison,Sarthou,
&Roger-Estrade,2010).
Organicfarmingat the field scale and landscape complexity(i.e.
theamountofsemi-naturalhabitatsinthelandscape)areamongthe
keymanagementoptionsfordiversifyingtheenvironmentandpoten-
tially limiting pest pressure(Bengtsson, Ahnström, & Weibull,2005;
Chaplin-Kramer etal., 2011). However, the relativeeffects of these
variablesatmultiplespatialscalesonpestcommunitiesandcroppro-
ductivityremainpoorlyexplored.Ithasbeenproposedthatlandscape
complexitymaynonlinearlymodifytheeffects oflocalfieldmanage-
ment on biodiversity and ecosystem services, such as pest control
(Concepción,Díaz,&Baquero,2008).Thissuggeststhatorganicfarm-
ingatthefieldscalewouldhaveamaximizedeffectin landscapesof
intermediatecomplexity,whileit wouldhavea minimaleffectin ex-
tremelysimplifiedorextremelycomplexlandscapes.However,studies
exploringthis hypothesisyielded contrastingresults(Birkhoferetal.,
2016;Winqvistetal.,2011).
In addition, most of the studies on pest ornatural enemy com-
munities have focused on the role of semi-natural habitats rather
thanthatoffarmingsystematthelandscapescale.Fewstudieshave
demonstratedthemulti-scale effectsoforganicfarmingon biodiver-
sitynor illustratedpotential interactionsbetween organicfarmingat
thelocalandlandscapelevels(Gabrieletal.,2010;Inclánetal.,2015).
However,howorganic farming at the landscape scale modulatesits
benefits at the local scale remains unclear. Moreover, most of the
studiesexaminingtheeffectsoforganicfarmingatthelandscapescale
focusonbiodiversityandlittle isknown aboutpest abundances(but
seeGosme,DeVillemandy,Bazot,&Jeuffroy,2012).Ontheonehand,
wecouldhypothesizethatagreaterproportionoforganicfarmingat
the landscape scale would promotethe diversity and abundance of
naturalenemies,enhancebiologicalcontrolandlimitpestabundance.
Onetheotherhand,wecouldpostulatethatfieldsunderorganicfarm-
ing may benefit fromreduced pest pressure at the landscape scale
owingtopesticideuseinlandscapeswithahighproportionofconven-
tionalfarming(“thechemicalumbrellaeffect”).Thus,scale-dependent
processes and the interplay between farming practices and semi-
naturalhabitatsonpestcommunitiesneedtobeinvestigated.
Despite the increased use ofpesticides world-wide, crop losses
owing to pests can still be substantial suggesting mixed effects of
pesticideoncropproductivity (Oerke,2006). Evidence ofthe coun-
terproductiveeffects ofpesticideuse have beenreportedin the lit-
erature(Bommarco,Miranda,Bylund,&Björkman,2011;Settleetal.,
1996),andhighlightthat,surprisingly,therelationshipsbetweenpest
pressure,pesticideuseandcropproductivityarepoorlydocumented.
Moreover, studies quantifying crop damage and productivity loss
owingtopestsmostlyconsidersinglespecies,althoughcropplantsare
attackedbymultiplespeciesthatcaninteract(Gagicetal.,2016).Thus,
itremainsdifficult to predict croplossesresultingfrom a largepest
community because (1) both synergistic and antagonisticeffects of
multiplepestattacksonplantperformanceexist(Stephens,Srivastava,
&Myers,2013);and(2)relationshipsbetweenpestcommunitiesand
cropdamagearehighlycontext-dependent(Savary,Teng,Willocquet,
&Nutter,2006).Itisthusofmajorimportancetoinvestigatehowpest
communitiesaffectcropproductivityifweintendtoreducepesticide
useinagroecosystems.
Inthis study,we investigatedtherelativeeffectsof farmingsys-
temsandsemi-naturalhabitatsatmultiplespatialscalesonpestcom-
munities(hereafter,pestsrefertobotharthropodpestsandpathogens)
andcropproductivitylevelsinvineyards.Weselectedviticultureas a
modelsystembecauseitishighlydependentonpesticidesandissub-
jectedto adiverse pestcommunity.Usinganexperimentaldesignin
whichpairsoforganicandconventionalfarmingfieldswereselected
alongtwoorthogonallandscapegradients(proportionofsemi-natural
habitatsand proportionoforganic farming),wewereableto unravel
therelative effectsofthesevariables onpest communities andcrop
productivity.Wehypothesizedthatpestinfestationswouldbegreater
inorganicthaninconventionalfieldsbecauseorganicfarmerscannot
usecurativepesticidesunlikeconventionalvinegrowers.Wealsohy-
pothesizedthatthelocaleffectsoforganicfarmingonpestinfestation
wouldbefurthermodulatedbythelandscapecontext.Weexpected,
on average,greater pest infestations in landscapes with a high pro-
portion of organic farming compared with landscapes containing a
highproportionofconventionalfarmingbutlowerpestinfestationsin
landscapeswitha highproportionofsemi-naturalhabitats compared
withlandscapescontainingalowproportionofsemi-naturalhabitats.
Finally,wehypothesizedthatcropproductivitywoulddirectlydepend
onthelevelofpestinfestation.
2 | MATERIALS AND METHODS
2.1 | Study sites and design
Our study sites were located within a vineyard-dominated region
(44°81′N, −0°14′W) of the Bordeaux area in southwestern France.
|
3
Journal of Applied Ecology
MUNERET ET al.
Thisis amajor wineproduction areawith ~138,000haof vineyards,
and it receives about 16 pesticide treatments a year per unit area
(Agreste, 2013). Our study system consisted of 21 pairs of organic
and conventional fields (42 fields) selected along two orthogonal
landscapegradients:oneofsemi-naturalhabitatsandone oforganic
farming (Table S1, Figure S1). The spatial scale used for calculating
landscapevariables and selectingsiteswas a1,000mradius around
each field. We only included organic vineyards that had been con-
vertedforatleast5years(onaverage11yearssincetheconversion;
TableS1). Thisdesignallowed fortheunravelling offarmingsystem
effectsat thelocalscale aswell astheeffects oftheproportions of
semi-naturalhabitatsandorganicfarmingatthelandscapescale.The
averagedistance betweenthetwo fieldsofa givenpairwas 125m.
In addition, landscape variables were calculated at three other spa-
tialscales:250-,500-and750-mradiiaroundeachfieldusingArcGIS
10.1(ESRI).Independenceamonglandscapevariableswasmaintained
atallscales.
2.2 | Studied pest taxa
Seven pest taxa, including five arthropod pests and two patho-
gens,wereregularlyquantifiedoverfourperiodsbetweenMayand
September2015. Allpesttaxa werecounted on30vine stocksat
each time period. We counted pests on four to six vine rows lo-
cated between the 5th and 15th closest vine rows of the paired
fields. Sampled vine stocks were more than 10m from the edge
or any other sampled vine stock. On each vine stock, the trunk,
threeleaves(thefirstoneatthehead,thesecondinthemiddleof
thevegetation and thethird at thebase)and threegrapeclusters
(randomly chosen) were carefully inspected in the field. On each
plant, the occurrence of mealybugs (Pseudococcidae) on trunks,
downymildew (Plasmopara viticola),black rot(Guignardia bidwellii),
mite galls (Colomerus vitis), phylloxera galls (Daktulosphaira vitifo-
liae)andleafhopperlarvae(Cicadellidae)onleaves,andlarvalnests
ofthegrape moths (Lobesia botrana and Empoecilia ambiguella)on
grape clusters were recorded. Most of the sampled leafhoppers
(>95%)belongedtothespeciesEmpoasca vitisandthe mostabun-
dant mealybug species were Parthenolecanium corni (>95%) and
Pulvinaria vitis(<5%).Botrytisbunchrot(Botrytis cinerea)andpow-
derymildew (Erysiphe necator) occurredonleaves and grapeclus-
ters,buttheoccurrencewassolowthattheywerenotincludedin
ouranalyses. Wesurveyed thesepestsbecausetheyarethemain
taxaattackinggrapevinesinourstudyregion.Downymildew,black
rot,leafhoppersandgrapemothsarethemajorpests,whilephyllox-
era,mitesandmealybugsareconsideredminorpests(Delièreetal.,
2016;Pertotetal.,2017).Noeconomicthresholdhasbeenidenti-
fiedformostofvineyardpestsbut ifmorethan5%oftheclusters
areattackedby grape moths or more than30%ofthe leaves are
attackedbypathogens,thenyieldlossishighlyprobableinthisre-
gion(Savary,Delbac, Rochas, Taisant, & Willocquet, 2009;Thiéry
&Moreau,2005;L.Delière, pers. comm.). In addition to pest oc-
currence and abundance, we calculated pest community richness
andevennessusingPielou’sindexatthefieldscaletoanalysehow
pestcommunities(sensulato)respondtofarmingsystemsandsemi-
naturalhabitats.Assumingthatorganicfarmingismorebeneficialto
peststhanconventionalfarming,weexpectedgreaterpestrichness
andevennesslevelsinorganiccomparedwithconventionalfields.
2.3 | Measurements of pesticide use intensity, crop
vigour and crop productivity
The42vineyardsweremanagedasusualby38differentfarmers.We
collecteddataonpesticideapplicationsforeachfieldbyinterviewing
thevinegrowersandcalculatingtheTreatmentFrequencyIndex(TFI)
forallpesticidesfollowingtheformula:
where n indicates the number of treatments, Appl. dose indicates
theapplied dose perhectareand Recom. doseindicatesthe recom-
mended dose per hectare (Haldberg, Verschuur, & Goodlass, 2005;
Jørgensenetal., 2008;OECD,2001). TheTFI isawell-knownindex
usedto assesspesticide pressureatdifferentscalesand tocompare
pesticide use intensity across different contexts (Jørgensen etal.,
2008).It is easytocalculate andallowsthe aggregation ofverydif-
ferentsubstancestomeasureoverall pesticidepressure.However,it
isnot anindexof toxicitybecauseit doesnot discriminate between
pesticideswithdifferentenvironmentaltoxicitylevels.Inthesurveyed
fields,86%ofthepesticidestargeteddownymildew andoidium,6%
wereinsecticidesagainstgrapemothsandleafhoppers,4%wereher-
bicidesandthe remaining were aimed at Botrytis(2%).Wealso as-
sessedcropvigourusingtheNormalizedDifferenceVegetationIndex
obtainedfrom ground-based measurementswithaGreenseekerleaf
colour analyser (N-Tech Industries, Ukiah, CA, USA and Oklahoma
StateUniversity,Stillwater,OK,USA)onceattheendofAugust2015
alonga 50-mtransect perfield.The meancropvigour perfieldwas
thendividedbythenumberofvinestocksalongthistransecttoelimi-
natetheeffectofvinestockdensityonthecropvigourscore.
Crop productivity was estimated a few days before harvest by
counting the number ofgrapes on 20 randomly chosen vine stocks
andbyweighing25randomlychosengrapesondifferentvinestocks.
Cropproductivityperhectarewascalculatedbymultiplyingtheaver-
agenumberofgrapespervinestockbytheaveragegrapeweightand
thevinestockdensityperfield.
2.4 | Statistical analyses
We used generalized linear mixed models (GLMMs) with Poisson
error distributions to investigate the effects of farming systems and
thelandscapecontextonthelevel of infestation for each pest taxa.
Theresponsevariablesusedinthemodelswerethenumberofleaves
ortrunksinfestedwithmealybugs,mites, downymildew, blackrot or
phylloxeraandthetotalnumberofleafhoppersorgrapemothscounted
perfield(n=166foreachpestexceptgrapemothsforwhichn=125).
Weusedamultimodel inferenceapproach totest ourhypotheses
andevaluatethesupportfromthedataforthreecompetingsetofmod-
elsofincreasingcomplexity.Foreachresponsevariable,westartedwith
TFI
=
∑n
i=1
Appl.dose
∕Recom.dose
,
4
|
Journal of Applied Ecology
MUNERET ET al.
afirstsetofmodels,M0,thatincludedfourexplanatoryvariables,“field
age”,“fieldsize”,“vinetrunkdensity”and“cropvigour”,whichwerecon-
sideredas potentialconfoundingvariables.Then,we selecteda setof
best models using the Akaikeinformation criteria corrected for small
samplesize(AICc).ModelsthatwerewithintherangeoftwoAICcunits
of the lowestAICc score were considered as the best set ofmodels
andwere usedto estimate themean effectsandconfidence intervals
ofeachpredictorvariableusingmodelaveraging(Grueber,Nakagawa,
Laws,& Jamieson,2011).Toaccountforthe studydesign, wealways
addedthe“fieldpairs”(21)andthe “samplingdate”(4)astwo crossed
randomeffects. GLMMswerecorrectedforoverdispersionby includ-
inganobservation-levelrandomeffect.Significantlocalcovariates(i.e.
witha high relativeimportance anda confidenceintervalsignificantly
differentfrom zero)retainedatthis step werethen used as thebasic
modelstructureinthetwoothersetsofcompetingmodels.Thefollow-
ingsetofmodels, M1, included previouslyselected“localcovariates”
and“local farmingsystem”asexplanatoryvariables.Thisstep enabled
thetestingofourhypothesisthatpestpopulationscouldbenefitfrom
organicfarmingatthelocalscale.Thelaststepofourmodellingproce-
dure,M2,includedsignificantlocalcovariatesselectedatM0,thelocal
farmingsystemandlandscapevariables(i.e.“theproportionoforganic
farming”and“theproportionofsemi-naturalhabitats”)atagivenscale.
Weaddedinteractiontermsbetweenthelocalfarmingsystemandboth
landscapevariablesinM2.Wedecidedtoalwaysuse“localfarmingsys-
tem”inM2asitallowedustotestourhypothesesonthemodulationof
theeffectsoflocalfarmingsystemsbylandscapecontext.Fourdifferent
setsofcompetingmodelswere consideredindependently usingland-
scapevariablescalculatedatfourdifferentspatialscales(250,500,750
and1,000m).Ateachstep (M0,M1and M2),weusedthesameaver-
agingapproachandthesamerandomstructureaspreviouslydescribed.
Foreveryresponsevariable andforeach top model ateachstep, we
calculated the marginal R2values and conditional R2valuestoassess
the amount ofvariance explained by the best model (i.e. that having
thelowestAICc; Nakagawa & Schielzeth,2013).FollowingSchielzeth
(2010),westandardizedallexplanatoryvariables,withmeanequalto0
andstandarddeviationequalto0.5beforemodelling.
Additionally,todeterminewhichlevelofmodelcomplexity,andin-
directlywhichspatialscale,wasthemostimportantforexplainingour
responsevariables, werecalculatedtheAkaikeweightsamongallof
themodelsfromthesixdifferentsets(i.e.M0,M1andM2atfourspa-
tialscales)obtainedforagivenresponsevariable.Usingthisapproach,
weestimatedtherelativeimportance ofeachlevelofcomplexityfor
agivenresponsevariable.ThesumoftheAkaikeweights(“SumWi”)
of the models obtained at a givenlevel of complexity provided the
model’sprobabilityofbeingatopmodelatallscales.
Thesame modellingstrategywas usedtoanalysehowpestrich-
ness,pestevenness,totalTFIandcropproductivityrespondedtoour
environmental variablesusing linear mixed models (LMMs) for pest
richnessandPoissonGLMMsfor other responsevariables (n=166,
n=166,n=42andn=38,respectively).Inaddition,weexaminedthe
effectsofpestinfestationsoncropproductivityusingLMMs.Theav-
eragepestinfestationsoftheseventaxawereincludedasexplanatory
variablesandthe“fieldpairs”asarandomfactor.
Diagnosticresidualplotsofallfullmodelswereconfirmedusingthe
DHARMapackage(Hartig,2017).Spatialautocorrelationintheresiduals
wereexploredusingvariograms,andnospatialautocorrelationwasde-
tected.Collinearityamong predictorswasassessedforeachfullmodel
usingthevarianceinflationfactor,andthevalueswereallcloseto1.
Allanalyseswereperformedusingrsoftware(RCoreTeam,2016)
andthepackages“lme4”(Bates,Mächler,Bolker,&Walker,2014)and
“MuMIn”(Bartoń,2016).
3 | RESULTS
3.1 | Relative effects of the farming system and
landscape context on each pest taxa
On average, 20.33% (SD=17.13), 4.7% (SD=9.35), 12.31%
(SD=9.45) and 4.47% (SD=7.18) of leaves were infested with
mites, phylloxera, black rot and downy mildew, respectively. On
average, 13.15% (SD=18.86) of trunks per field were infested
with mealybugs, 8.44% (SD=9.33) with leafhoppers and 1.72%
(SD=3.19) with grape moths. Approximately, 40% of the fields
were subjected to a level of pest attack that could lead to yield
loss.At thepopulationlevel, wefoundthat onlymealybugs, phyl-
loxera and mites responded to farming systems and semi-natural
habitats at multiple scales. Black rot, downy mildew, leafhoppers
andgrapemothsdidnotrespondtoanylocalorlandscapevariables
(seeTablesS2–S5).Amongthelocalcovariates,thecropvigouras
assessedbytheNormalizedDifferenceVegetationIndexwasnever
retainedasasignificantvariable.Atthelandscapescale,thepropor-
tionoforganicfarmingwasneverretainedasasignificantvariable
explainingpestinfestations.
3.1.1 | Mealybugs
Models including local covariates, the local farming system and
landscape variables had the highest probability of being among the
bestsets ofmodels(Table1). Inparticular,models fittedusingland-
scapevariables at the250-m scalehadthe highest probability(Sum
WiM2at250m=0.41)toappearas topmodelsamongallmodels fitted
atallscales. Modelaveragingof modelsfittedat thisspatialscalein-
dicatedthatlocalfarmingsystem, vinestockdensity,andproportion
ofsemi-naturalhabitatswereallincludedinthetopmodels(eachrela-
tivevariable’simportancewasequalto1;Table1).Mealybuginfesta-
tionwasgreaterin organic fields, increased with vinetrunkdensity
anddecreasedwiththeproportionofsemi-naturalhabitats(
R2
m
=0.15;
R2
c
=0.62;Table1).Wedidnotfindanysignificantinteractionsamong
thelocal farmingsystem andlandscape variables(Figure1a).Results
ofthemultimodel inferencesatother spatial scaleswereconsistent
withtheseresults.
3.1.2 | Mites
Models fitted using local covariates, the local farming sys-
tem and landscape variables at the 250-m scale had the highest
|
5
Journal of Applied Ecology
MUNERET ET al.
probabilityofbeingtopmodelsamongallmodelsfittedatallscales(Sum
WiM2at 250m=0.99). Model averaging at this scale showed that the
proportionofleavesinfestedwith mitegallsincreasedwithfield age
and decreased with trunk density, field size and the proportion of
semi-naturalhabitatsinthelandscape(eachrelativevariable’simpor-
tancewas equal to1;
R2
m
=0.44;
R2
c
=0.87;Table2). This modelalso
revealed a significant interaction between the local farming system
andthe proportionof semi-naturalhabitats.Thisindicated thatmite
infestationsweregreaterinconventionalfieldsthanin organicfields
in landscapes with a low proportion of semi-natural habitats, while
therewasnodifferenceinmiteinfestationlevelsbetweenorganicand
conventional farming in landscapes with a high proportion of semi-
naturalhabitats(Figure1b,Table2).
3.1.3 | Phylloxera
Models fitted using local covariates, the local farming system and
landscapevariablescalculated at the 1,000-mscalehad the highest
probability (Sum WiM2 at 1,000m=0.98) of being top models among
all models fitted at all scales. Model averaging of models fitted at
1,000m indicated that phylloxera infestations decreased with field
age and that the effect of the local farming system on phylloxera
infestation was dependent of the proportion of semi-natural habi-
tats (each relative variable’s importance was equal to 1;
R2
m
=0.25;
R2
c
=0.72;Table3).Thisinteractionindicatedthatthe level of phyl-
loxerainfestationwasgreater in organic fields than in conventional
fieldsbutonlyin landscapes with a high proportion of semi-natural
TABLE1 Modelselectiontableformodelsexplainingmealybuginfestationsinvineyards.Thetablereportstheexplanatoryvariables
selected,estimates,confidenceintervals(2.5%–97.5%)andtherelativeimportanceofeachlevelofmodelcomplexity(M0,M1andM2).M0
onlyconsideredlocalcofoundingvariables;M1consideredtheretainedlocalcovariatesfromM0aswellasthelocalfarmingsystem(organicor
conventional);andM2considerspreviousvariablesaswellaslandscapevariables.ForM2,onlymodeloutputsofthemostimportantspatial
scale(identifiedbythesumofAkaikeweightsnormalizedacrosseachspatialscale)areindicatedinthetable.Foreachlevelofmodel
complexity,R²marginalandR²conditionalarereported.R2valueswerecalculatedusingthebestmodelsateachscale.ThesumoftheAkaike
weightnormalizedacrosseachspatialscale(SumWi)providedtheprobabilityofagivenlevelofcomplexitytoappearinthetopmodels
Models Explanatory variables selected Estimates Confidence intervals
Relative variable
importance
M0(
R2
m
=0.06;
R2
c
=0.6;
sumWi<0.01)
Fieldage −0.33 −0.83to0.01 0.81
Vinetrunkdensity 0.86 0.29–1.42 1
Fieldsize 0.08 −0.15to0.74 0.28
Cropvigour −0.04 −0.73to0.31 0.17
M1(
R2
m
=0.08;
R2
c
=0.61;
sumWi=0.03)
Vinetrunkdensity 0.72 0.15–1.29 1a
Localfarmingsystem:Conventional 0.46 0.14–0.78 1a
M2at250m(
R2
m
=0.15;
R2
c
=0.62;sumWi=0.41)
Localfarmingsystem 0.49 0.15–0.84 1
Vinetrunkdensity 0.72 0.18–1.26 1
Proportionofsemi-naturalhabitats −0.8 −1.49to−0.11 1
Proportionoforganicfarming 0.16 −0.35to1.25 0.36
aOnlyonebestmodelselectedhere.
FIGURE1 Responsesofmealybugs(a),mites(b)andphylloxera(c)totheinteractionsbetweenlocalfarmingsystemsandtheproportionof
semi-naturalhabitatsattwodifferentspatialscales(250mor1,000m)
0
10
20
00.1 0.20.3 0.4
Proportion of semi−natural habitats
at 250 m
Number of tr
unks infested with mealybugs
(a)
0
20
40
60
00.1 0.20.3 0.4
Proportion of semi−natural habitats
at 250 m
Number of leaves infested with mites
(b)
0
20
40
60
00.250.5 0.75
Proportion of semi−natural habitats
at 1000 m
Number of le
aves infested with phylloxera
Farming systems
Conventional
Organic
(c)
6
|
Journal of Applied Ecology
MUNERET ET al.
habitats.Theoppositewasfoundinlandscapeswithalowproportion
ofsemi-naturalhabitats(Figure1c).
3.2 | Relative effects of the farming system and
semi- natural habitats on the pest community
Forpestcommunityevenness,models fittedwithlandscapevariables
calculatedatthe250-mscalehadthehighestcumulatedprobabilityof
beingtop models(SumWiM2 at250m=0.45; TableS6).Field age,vine
trunkdensityandtheproportionofsemi-naturalhabitatswereincluded
inthetopmodelsfittedatthe250-mscale(TableS6).Pestcommunity
evenness increased with vine trunk density and with the proportion
ofsemi-naturalhabitatsinthelandscape,whileitdecreasedwithfield
age (each relative variable’s importance was equal to 1;
R2
m
= 0.19;
R2
c
=0.32; Table S6). Our multimodel inference approach showed no
evidenceof any effectscausedby otherlocalor landscapevariables.
Additionally,therewerenoeffectsofthefarmingsystemandsemi-
naturalhabitatsatmultiplespatialscalesonthepesttaxarichness.
3.3 | Crop productivity and management intensity
Cropproductivitywasnotaffectedbythelocalcovariates,localfarm-
ingsystemoranylandscapevariables.Onaverage,organicfieldspro-
duced 11.01t (SD=4.07) of grape per hectare, while conventional
fieldsproduced onaverage 10.90t(SD=3.58) ofgrapeperhectare,
whichhighlighted thesimilar productionlevelsbetween organicand
conventionalsystems (Figure2a).Surprisingly, cropproductivitywas
notaffectedbyanypest infestationsof anytaxadespitevariabilities
inpestinfestationsamongfields(TableS7).Finally,thetotalTFIwas
lowerin organicthanin conventionalfieldsbut wasnot affectedby
thelandscapecompositionatanyscale.Onaverage,thetotalTFIwas
twofoldlowerinorganicthaninconventionalfields(Figure2b).
4 | DISCUSSION
Toourknowledge,thisisthefirststudyinvestigatinghowinfestations
with multiple pest taxa respond to both organic farming and semi-
naturalhabitatsusingamultiplescaledesignwithorthogonallandscape
factors.Contrarytoourinitialhypotheses,thepestcommunitydidnot
differatthefieldscalebetweenorganicandconventionalfarmingsys-
tems.Additionally, theproportion oforganic farminginthelandscape
wasneveranimportantvariableforexplainingpestinfestations.Finally,
ourstudyshowed that threeoutof seven pestspeciesresponded to
local farming systems and semi-natural habitats in the landscape.
Nevertheless,itillustratedtheimportanceofconsideringbothfarming
practicesandlandscapecontextwhenstudyingpestdynamics.
Contraryto expectations, the pestcommunitydid not differbe-
tweenorganicandconventionalfields.Additionally,increasingtheor-
ganicfarmingareainthelandscape(withintherangeexploredinour
studydesign)didnotleadtogreaterpestinfestations.Ourresultscon-
tradictpreviousresultsfromtwotheoreticalstudiesinvestigatingthe
effectsofthespatialarrangementoforganicfarminginthelandscape
(Adl, Iron, & Kolokolnikov,2011; Bianchi, Ives, & Schellhorn, 2013).
Bothstudies showed thatthedeployment oforganicfarming at the
landscapescalecouldincreasepestoutbreaks.However,thesestudies
assumedthatchemical controlwas lesseffective,orevenabsent, in
organicfieldscomparedwithconventionalones,whichisnotthecase
TABLE2 Modelselectiontableformodelsexplainingmiteinfestationsinvineyards.Allmodels,explanatoryvariables,estimates,confidence
intervalsandrelativeimportancesthatarereportedinthistablehavebeenobtainedusingthesameprocedureasthedatareportedinthe
Table1.SeethelegendinTable1
Models Explanatory variables selected Estimates Confidence intervals
Relative variable
importance
M0(
R2
m
= 0.17;
R2
c
=0.63;
sumWi<0.01)
Fieldage 0.81 0.51–1.10 1
Vinetrunkdensity −0.41 −0.74to−0.08 1
Fieldsize 0.47 −0.77to−0.18 1
Cropvigour 0.03 −0.22to0.45 0.3
M1(
R2
m
= 0.19;
R2
c
=0.65;
sumWi<0.01)
Fieldage 0.63 0.33–0.92 1
Localfarmingsystem:Conventional −0.44 −0.64to−0.23 1
Vinetrunkdensity −0.19 −0.61to0.02 0.63
Fieldsize −0.4 −0.68to−0.11 1
M2at250m(
R2
m
=0.44;
R2
c
=0.87;sumWi=0.99)
Vinetrunkdensity −0.38 −0.69to−0.13 1
Fieldage 0.69 0.48–0.98 1
Fieldsize −0.5 −0.79to−0.28 1
Localfarmingsystem:Conventional −0.19 −0.36to0.02 1
Proportionofsemi-naturalhabitats −1.48 −2.04to−1.07 1
Proportionoforganicfarming −0.08 −0.72to0.42 0.27
Localfarmingsystem:proportionof
semi-naturalhabitats
0.5 0.22–0.95 1
|
7
Journal of Applied Ecology
MUNERET ET al.
here. Oursurvey of farming practices showed that organic growers
wereusingorganicpesticidestocontrolpests,asindicatedbyanaver-
agetotalTFIofapproximately9fororganicfields,whiletheyobtained
similarlevelsofcropproductivity.Furthermore,ourresultsareinline
withtwoempirical studies thatshowedno evidenceofgreaterpest
pressure in cerealfields and orchards as the amount of area under
organicfarmingin thelandscape increased(Gosmeetal.,2012;Ricci
etal.,2009).
Threeoutofthesevensurveyedpesttaxarespondedtothepro-
portion of semi-natural habitats in the landscape. Both mealybug
andmiteinfestations decreased astheproportion of semi-natural
habitatsinthelandscapeincreased.Similarly,thephylloxeraabun-
dancedecreasedwiththeproportionofsemi-naturalhabitatsinthe
landscape but only in conventional fields.This negative effect of
landscapecomplexityonpestinfestationsmaybeexplainedbytwo
complementaryhypotheses:(1)directeffectsoflandscapecompo-
sitiononpestdynamicsand(2)indirecteffectsoflandscapecompo-
sitiononpeststhroughtop-downcontrolbynaturalenemies(Rusch
etal.,2010;Veres,Petit,Conord,&Lavigne,2013).Directeffectsof
landscapecomplexityonpestdynamicsincludelimitedpestdisper-
salowingto the direct barriereffectsof unsuitable habitattypes,
such as semi-natural habitats, and reduced pest sources in more
complex landscapes that support lower proportionsof host plant
(Avelino,Romero-Gurdián,Cruz-Cuellar,&Declerck,2012;Kuefler,
Hudgens,Haddad,Morris,&Thurgate,2010;Plantegenest,LeMay,
& Fabre,2007; Summerville, 2004). Indirect effects of landscape
complexityon pests are caused by the presence of keyresources
fornaturalenemies(Landis,Wratten,&Gurr,2000).Theincreased
availabilityofsemi-natural habitatsinthe landscape enhancesthe
diversityand abundance of natural enemies and, in turn, the bio-
logical control of pests (Chaplin-Kramer etal., 2011; Letourneau,
Jedlicka,Bothwell,&Moreno,2009;Ruschetal.,2016).
Mealybugs, mites and phylloxera were also affected by local
farmingsystems,either aloneorthroughinteractionswiththe pro-
portionofsemi-natural habitats in thelandscape.Mitesandphyl-
loxerawerethetwotaxa thatsupported ourhypothesis predicting
that the effects of the local farming system on pest infestation
wouldbemodulatedby the proportion ofsemi-natural habitatsin
thelandscape.Invery simple landscapes, mites and phylloxerain-
festationswere lower inorganicthaninconventional fields, while
incomplex landscapes,theinfestation levelswereeithersimilaror
greaterinorganicthaninconventionalfields.Thisillustrateshowthe
landscapecontextcanmodulatetheeffectoflocalfarmingsystems
onpestinfestations.However,ourresultsseem tonotcorroborate
theintermediatelandscape complexityhypothesisthat statedthat
organic farming would be more effective in enhancing ecosystem
services,suchasbiologicalcontrol,inintermediatelandscapesthan
in simple and complex landscapes (Tscharntke, Tylianakis, etal.,
2012).Phylloxerainfestationsweregreaterinorganicfieldsthanin
conventionalfieldsbutonlyincomplexlandscapes.Thus,processes
otherthantop-downcontrolbynaturalenemiesmightbeinvolved
and other covariatesrelated to bottom-up processes mayexplain
this pattern. Despite the negative effect ofsemi-natural habitats,
TABLE3 Modelselectiontableformodelsexplainingphylloxerainfestationsinvineyards.Allmodels,explanatoryvariables,estimates,
confidenceintervalsandrelativeimportancesthatarereportedinthistablehavebeenobtainedusingthesameprocedureasthedatareported
intheTable1.SeethelegendinTable1
Models Explanatory variables selected Estimates Confidence intervals
Relative variable
importance
M0(
R2
m
= 0.07;
R2
c
=0.66;
sumWi<0.01)
Fieldage −1.11 −1.74to−0.46 1
Cropvigour −0.22 −0.99to0.13 0.51
Vinetrunkdensity −0.08 −1.15to0.52 0.24
Fieldsize −0.06 −0.88to0.37 0.24
M1(
R2
m
= 0.07;
R2
c
=0.66;
sumWi<0.01)
Fieldage −1.11 −1.76to−0.47 1
Localfarmingsystem:Conventional −0.04 −0.54to0.25 0.31
M2at1,000m(
R2
m
=0.25;
R2
c
=0.72;sumWi=0.98)
Fieldage −1.12 −1.54to−0.67 1
Localfarmingsystem:Conventional −0.36 −0.63to−0.10 1
Proportionoforganicfarming 0.71 −0.09to2.51 0.58
Proportionofsemi-naturalhabitats 0.72 −0.34to1.79 1
Localfarmingsystem:proportionof
semi-naturalhabitats
1.82 1.37–2.27 1
FIGURE2 Effectsoflocalfarmingsystems(conventionalor
organic)on(a)cropproductivity(t/ha)and(b)thetotalTreatment
FrequencyIndex(TFI)
5
10
15
ConventionalOrganic
Crop productivity (tons/ha)
(a)
10
20
30
Conventional Organic
Total TFI
(b)
Farming systems
8
|
Journal of Applied Ecology
MUNERET ET al.
infestations by mealybugswere always greater in organic than in
conventionalvineyardssuggestingthatorganic vineyardscouldbe
highlybeneficial tomealybugs.Indeed, mealybugsarefavouredby
thetillage,thepresenceofFormicidaeandadecreasedexposureto
syntheticpesticides(Daaneetal.,2012;Mgocheki&Addison,2010;
Muscasetal.,2017).
Pesttaxathat respondedtothe landscapecontextwereaffected
by the proportion of semi-natural habitats at two different spatial
scales:the250-m(formealybugsandmites)andthe1,000-mradii(for
phylloxera).The most important scale explaining species abundance
canbeinterpretedasthescaleatwhichagivenspeciesperceivesand
interactswiththelandscapeandcanbeusedformanagementissues
(Jackson & Fahrig,2012). This scale of response depends on func-
tionalattributes,suchas the dispersal ability, ofthe species and on
thestructureofthelandscapeitself.Themostimportantscalesfound
in our study forthese three pest taxa are consistent with available
informationon their dispersal abilities (Forneck,Anhalt, Mammerler,
& Griesser, 2015; Grasswitz &James, 2008). Moreover, ourresults
corroboratethose of recent studies in perennial crops in which the
beneficialeffectsofsemi-naturalhabitatsonpestcontrolweregreat-
estatrelativelysmallspatialscales(<250mradius)(Henrietal.,2015;
Thomson&Hoffmann,2013).
Pestpressureatthecommunityleveltendstobelimitedin com-
plex landscapes mainly owing to the negative effect of landscape
complexityonmites.Downymildew,blackrot,leafhoppersandgrape
mothswere not affected bytheproportion of semi-naturalhabitats
inthelandscape.Overall,these findingsareconsistentwiththecon-
clusionsoftworecentmeta-analysesthatfoundnocleartrendinthe
responseof pest abundancetothe proportion ofsemi-natural habi-
tatsinthelandscape,despitethestrongpositiveeffectsoflandscape
complexityonnaturalenemiesandbiologicalcontrol(Chaplin-Kramer
etal.,2011;Veresetal.,2013).Onepossibleexplanationisthatfarm-
ingpractices,andpesticideuseinparticular,mayhavehiddentheef-
fectoflandscapecontextonpestpopulations(Tscharntkeetal.,2016;
Veresetal., 2013). Indeed,the fourpest taxa thatdid not showany
responsetofarmingpracticesorlandscapecontextarethemaintar-
getsofpesticideuseinbothorganicandconventionalvineyards.This
strongly suggests that both farming systemsused effective control
strategiesthatmaskedpotentialeffects.Second,thethematicresolu-
tion,aswellasthespatialextent,usedtocharacterizeourlandscapes
mightdifferfromthe actualfunctionalrolesofthehabitatsandfrom
the scale of response of the given pest species (Jackson & Fahrig,
2012). The most relevant scale of observation corresponds to the
meandispersal distanceof thespecies beingstudied(Gilligan,2008;
Jackson & Fahrig,2012). Thus, the lack of landscape contexteffect
forthefourmajorpestsmightalsoresultfroma toonarrowscaleof
observation,especiallyforpathogensthatdispersebywindoverlong
distances(Fontaineetal.,2013).
Finally, ourstudy demonstrated that organic vineyards haveTFIs
that are twofoldless important than those of conventionalvineyards
havingsimilar levelsof pestinfestations andequalproductivitylevels.
TheTFIisclearlya proxyfortreatmentintensityandisnotameasure
of environmentalimpact. However, increasingthe amount of organic
farming at the landscape scale should decrease treatment intensity
withoutmodifyingpestpressureorcropperformanceinthegivencon-
textofourstudy.Ourresultsareconsistentwitharecentstudyonarable
cropsinwhichthereductionofpesticideusedidnotaffectcropproduc-
tivityoreconomicperformances(Lechenet,Dessaint,Py,Makowski,&
Munier-Jolain,2017).Studiesregardingtheagronomicconsequencesof
organicfarmingexpansionshouldalsoconsidercropquality.
4.1 | Synthesis and applications
Reducing pesticide use, while not altering crop productivity and
quality,isa major challenge foragroecologists.Our study demon-
stratesthatorganic farming can be used asanagri-environmental
schemetoachieve thisgoalinvineyardagroecosystemsthatheav-
ily depend on synthetic pesticides. Increasing the area of organic
farmingin the landscapedidnot leadtogreater pestpressurebut
reduced treatment intensity and maintained crop productivity.
However,thelong-termeffectsoforganicinputslargelyusedinor-
ganicfarmingsuchassulphurandcopper shouldbeinvestigatedin
vineyards.Moreover,ourresultsfurtherillustratehowproportions
ofsemi-naturalhabitatsinthelandscapecanmodulatethelocalef-
fectsof farmingsystemson pestinfestationsby sometaxa.These
results highlight the importance of taking into account landscape
composition to optimize farming system allocation and limit pest
pressure.Ourstudyhasimportant implicationsforbothpractition-
ersand policymakersconcernedwith theecologicalintensification
offarmingsystemsandland-useplanning.However,futureresearch
is still needed to explore potential threshold effects when there
is a much greater proportion of organic farming in the landscape.
Moreover,the relative effectsofthe spatialcompositionand con-
figuration(e.g.connectivity)oforganicfarmingonpestpressureand
biologicalpest controlremain largelyunknownandrequirefurther
investigation.
ACKNOWLEDGEMENTS
Weare gratefultoArthur Auriol,EmilieVergnes, LauraArias, Lionel
Druelle, Pascale Roux, Olivier Bonnard, Sylvie Richard-Cervera,
IsabelleDemeaux, LisaLe Postec,LionelDelbacandGillesTarrisfor
theirtechnical help.Wealsothankthe 38grapevinegrowersforal-
lowing us access to their vineyards and Lesley Benyon, PhD, from
EdanzGroupfor editingadraft ofthismanuscript. Theresearchwas
funded by the Region Aquitaine (REGUL project) and the Agence
Française pour la Biodiversité, joint call Ecophyto & the French
National Foundation for Research on Biodiversity (SOLUTION pro-
ject).ThisresearchisalsopartoftheclusterofExcellenceCote.
AUTHORS’ CONTRIBUTIONS
L.M.,D.T.andA.R.conceivedtheworkanddesignedtheexperiments;
L.M.,B.J.andA.R.collectedthedata;L.M.andA.R.analysedthedata;
L.M.and A.R. ledthewriting ofthemanuscript. All authorscontrib-
utedcriticallytothedraftsandgavefinalapprovalforpublication.
|
9
Journal of Applied Ecology
MUNERET ET al.
DATA ACCESSIBILITY
Data available from the Dryad Digital Repository https://doi.
org/10.5061/dryad.vv1t9(Muneret,Thiéry,Joubard,&Rusch,2017).
ORCID
Adrien Rusch http://orcid.org/0000-0002-3921-9750
REFERENCES
Adl,S.,Iron,D.,&Kolokolnikov,T.(2011).Athresholdarearatiooforganic
to conventional agriculture causes recurrent pathogen outbreaks in
organicagriculture.Science of the Total Environment,409,2192–2197.
Agreste.(2013). Ministèrede l’agriculture,de l’agroalimentaireetde la
forêt–Agreste–Lastatistique,l’évaluationetlaprospectiveagricole–
Pratiquesculturales dansla viticulture.http://agreste.agriculture.go
uv.fr/enquetes/pratiques-culturales/pratiques-culturales-dans-la/.
Avelino, J.,Romero-Gurdián, A., Cruz-Cuellar, H. F., & Declerck, F.A.
(2012).Landscapecontextandscaledifferentiallyimpactcoffeeleaf
rust,coffeeberryborer,andcoffeeroot-knotnematodes.Ecological
Applications,22,584–596.
Bartoń,K.(2016).Package‘MuMIn’.Modelselectionandmodelaverag-
ingbasedoninformationcriteria.Rpackageversion1.15.11.http://
CRAN.R-project.org/package=MuMIn
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2014). Fitting linear
mixed-effectsmodelsusinglme4.arXivpreprintarXiv:1406.5823.
Bengtsson,J., Ahnström,J., &Weibull,A.-C.(2005).The effectsofor-
ganic agriculture on biodiversity and abundance:A meta-analysis.
Journal of Applied Ecology,42,261–269.
Bianchi,F.J.J.,Ives,A. R., &Schellhorn,N. A.(2013). Interactionsbe-
tween conventional and organic farming for biocontrol services
acrossthelandscape.Ecological Applications,23,1531–1543.
Birkhofer, K.,Arvidsson, F., Ehlers, D., Mader, V. L., Bengtsson, J., &
Smith, H. G. (2016). Organicfarming affects the biological control
ofhemipteranpestsandyieldsinspringbarleyindependentofland-
scapecomplexity.Landscape Ecology,31,567–579.
Bommarco,R.,Miranda,F.,Bylund,H.,&Björkman,C.(2011).Insecticides
suppress natural enemies and increase pest damage in cabbage.
Journal of Economic Entomology,104,782–791.
Chaplin-Kramer,R.,O’Rourke,M.E.,Blitzer,E.J.,&Kremen,C.(2011).A
meta-analysisofcroppestandnaturalenemyresponsetolandscape
complexity.Ecology Letters,14,922–932.
Concepción,E.D.,Díaz,M.,&Baquero,R.A.(2008).Effectsoflandscape
complexity on the ecological effectiveness of agri-environment
schemes.Landscape Ecology,23,135–148.
Daane,K.M., Almeida, R. P.P.,Bell,V.A., Walker,J.T.S.,Botton,M.,
Fallahzadeh,M., … Zaviezo,T. (2012).Biologyand management of
MealybugsinVineyards.Arthropod Management in Vineyards(edsN.
J.Bostanian,C. Vincent & R.Isaacs),pp. 271–307. Dordrecht,the
Netherlands:Springer.
Delière,L.,Burgun,X.,Lafond,D.,Mahé, H.,Métral, R.,Serrano, E.,…
Pillet, E. (2016). Réseau DEPHY EXPE: Synthèse des résultats à
mi-parcoursà l’échellenationale - filièreViticulture.Celluled’Ani-
mationNationale DEPHYEcophyto. http://viticulture.ecophytopic.
fr/sites/default/files/Synth%C3%A8se%20EXPE%20Viticulture.
pdf.
Foley,J. A., Ramankutty, N., Brauman,K. A., Cassidy, E. S., Gerber, J.
S.,Johnston,M.,…Zaks,D.P. M.(2011).Solutionsforacultivated
planet.Nature,478,337–342.
Fontaine,M.C.,Austerlitz,F.,Giraud,T.,Labbé,F.,Papura,D., Richard-
Cervera,S.,& Delmotte,F.(2013).Genetic signatureofarangeex-
pansionand leap-frogeventaftertherecentinvasion ofEuropeby
thegrapevinedownymildewpathogenPlasmopara viticola. Molecular
Ecology,22,2771–2786.
Forneck,A.,Anhalt,U.C.,Mammerler,R.,&Griesser,M.(2015).Noev-
idenceofsuperclones inleaf-feedingformsofAustriangrapephyl-
loxera(Daktulosphaira vitifoliae).European Journal of Plant Pathology,
142,441–448.
Gabriel, D., Sait, S. M., Hodgson, J.A., Schmutz, U., Kunin, W. E., &
Benton,T.G.(2010).Scalematters:Theimpactoforganicfarmingon
biodiversityatdifferentspatialscales.Ecology Letters,13,858–869.
Gagic,V.,Riggi,L.G.,Ekbom,B.,Malsher,G.,Rusch,A.,&Bommarco,R.
(2016).Interactive effectsofpestsincreaseseedyield.Ecology and
Evolution,6,2149–2158.
Gilligan, C. A. (2008). Sustainable agriculture and plant diseases: An
epidemiological perspective. Philosophical Transactions of the Royal
Society of London B: Biological Sciences,363,741–759.
Gosme, M., De Villemandy, M., Bazot, M., & Jeuffroy, M.-H. (2012).
Localandneighbourhoodeffectsoforganicandconventionalwheat
management on aphids, weeds, and foliar diseases. Agriculture,
Ecosystems & Environment,161,121–129.
Grasswitz,T.R., &James,D. G.(2008). Movementofgrapemealybug,
Pseudococcus maritimus, on and between host plants. Entomologia
Experimentalis et Applicata,129,268–275.
Grueber, C. E., Nakagawa, S., Laws, R. J., & Jamieson, I. G. (2011).
Multimodelinferenceinecologyandevolution:Challengesandsolu-
tions.Journal of Evolutionary Biology,24,699–711.
Haldberg,N.,Verschuur,G.,& Goodlass,G. (2005).Farmlevel environ-
mentalindicators;aretheyuseful?Anoverviewofgreenaccounting
systemsfor Europeanfarms. Agriculture, Ecosystems & Environment,
105,195–212.
Hartig,F.(2017).DHARMa:Residualdiagnosticsforhierarchical(multi-
level/mixed) regression models. https://cran.r-project.org/web/
packages/DHARMa/index.html.
Henri,D.C.,Jones,O.,Tsiattalos,A.,Thébault,E.,Seymour,C.L.,&Veen,
F.J. (2015). Natural vegetation benefits synergisticcontrol of the
three main insect and pathogen pests of a fruit cropin southern
Africa.Journal of Applied Ecology,52,1092–1101.
Inclán,D.J.,Cerretti,P.,Gabriel,D.,Benton,T.G.,Sait,S.M.,Kunin,W.
E.,…Marini, L. (2015).Organicfarming enhancesparasitoiddiver-
sityatthelocal andlandscapescales. Journal of Applied Ecology,52,
1102–1109.
Jackson,H. B., & Fahrig,L. (2012). What sizeisa biologically relevant
landscape?Landscape Ecology,27,929–941.
Jørgensen, L. N., Noe, E., Nielsen, G. C.,Jensen, J. E., Ørum, J. E., &
Pinnschmidt,H.O.(2008).Problemswithdisseminatinginformation
ondiseasecontrolinwheatandbarleytofarmers.European Jornal of
Plant Pathology,121,303–312.
Kuefler,D., Hudgens,B., Haddad, N. M.,Morris,W.F.,&Thurgate, N.
(2010).Theconflictingroleofmatrixhabitatsasconduitsandbarri-
ersfordispersal.Ecology,91,944–950.
Landis,D.A.,Wratten,S.D.,&Gurr,G.M.(2000).Habitatmanagement
toconservenaturalenemiesofarthropodpestsinagriculture.Annual
Review of Entomology,45,175–201.
Lechenet, M., Dessaint, F., Py, G., Makowski, D., & Munier-Jolain, N.
(2017). Reducing pesticide use while preserving cropproductivity
andprofitabilityonarablefarms.Nature Plants,3,17008.
Letourneau,D.K.,Armbrecht,I.,Rivera,B.S.,Lerma,J.M.,Carmona,E.J.,
Daza,M.C.,…Trujillo,A.R.(2011).Doesplantdiversitybenefitagro-
ecosystems?Asyntheticreview.Ecological Applications,21,9–21.
Letourneau,D.K.,Jedlicka,J.A.,Bothwell,S.G.,&Moreno,C.R.(2009).
Effectsofnatural enemybiodiversityon thesuppression ofarthro-
pod herbivoresin terrestrial ecosystems. Annual Review of Ecology,
Evolution, and Systematics,40,573–592.
Mgocheki, N., & Addison, P. (2010). Spatial distribution of ants
(Hymenoptera:Formicidae),vine mealybugd andmealybugparasit-
oidsinvineyards.Journal of Applied Entomoly,134,285–295.
10
|
Journal of Applied Ecology
MUNERET ET al.
Muneret, L., Thiéry, D., Joubard, B., & Rusch, A. (2017). Data from:
Deploymentof organicfarmingat a landscapescalemaintains low
pestinfestationandhighcropproductivitylevelsinvineyards.Dryad
Digital Repository,https://doi.org/10.5061/dryad.vv1t9
Muscas,E.,Cocco,A.,Mercenaro,L.,Cabras,M.,Lentini,A.,Porqueddu,
C., & Nieddu, G. (2017). Effects ofvineyard floor cover crops on
grapevine vigor,yield, and fruit quality, and the development of
the vine mealybug under a Mediterranean climate. Agriculture,
Ecosystems & Environment,237,203–212.
Nakagawa,S.,&Schielzeth,H.(2013).Ageneralandsimplemethod for
obtainingR2fromgeneralizedlinearmixed-effectsmodels.Methods
in Ecology and Evolution,4,133–142.
OECD.(2001).Environmentalindicatorsforagriculturevolume3/2001.
https://www.oecd.org/tad/sustainable-agriculture/40680869.pdf.
Oerke, E.-C. (2006). Crop losses to pests. The Journal of Agricultural
Science,144,31–43.
Pertot, I., Caffi,T., Rossi,V., Mugnai, L., Hoffmann, C., Grando, M. S.,
…Anfora, G. (2017). A critical reviewof plant protection tools for
reducingpesticideuse on grapevine and new perspectivesforthe
implementationofIPMinviticulture.Crop Protection,97,70–84.
Plantegenest,M.,LeMay,C.,&Fabre,F.(2007).Landscapeepidemiologyof
plantdiseases.Journal of the Royal Society Interface,4,963–972.
RCoreTeam.(2016).R:TheRprojectforstatisticalcomputing.https://
www.r-project.org/.
Ricci,B.,Franck,P.,Toubon,J.F.,Bouvier,J.C.,Sauphanor,B.,&Lavigne,
C.(2009).Theinfluenceoflandscapeoninsectpestdynamics:Acase
studyinsoutheasternFrance.Landscape Ecology,24,337–349.
Rusch, A., Chaplin-Kramer, R., Gardiner, M. M., Hawro, V., Holland,
J., Landis, D., … Bommarco,R. (2016). Agricultural landscape sim-
plification reduces natural pest control: A quantitative synthesis.
Agriculture, Ecosystems & Environment,221,198–204.
Rusch, A., Valantin-Morison, M., Sarthou, J.-P., & Roger-Estrade, J.
(2010).Biologicalcontrolofinsectpestsinagroecosystems:Effects
ofcrop management,farmingsystems, andseminaturalhabitats at
thelandscapescale:Areview.Advances in Agronomy,109,219–260.
Savary,S., Delbac, L., Rochas, A., Taisant, G., & Willocquet, L. (2009).
Analysisofnonlinearrelationshipsindualepidemics,anditsapplica-
tiontothemanagementofgrapevinedownyandpowderymildews.
Phytopathology,99,930–942.
Savary, S., Teng, P. S., Willocquet, L., & Nutter, F. W. Jr (2006).
Quantification and modeling ofcrop losses: A reviewof purposes.
Annual Review of Phytopathology,44,89–112.
Schielzeth,H.(2010).Simplemeanstoimprovetheinterpretabilityofre-
gressioncoefficients.Methods in Ecology and Evolution,1,103–113.
Settle, W. H., Ariawan, H., Astuti, E. T., Cahyana, W., Hakim, A. L.,
Hindayana,D., &Lestari,A.S. (1996). Managingtropical ricepests
throughconservation ofgeneralist natural enemiesand alternative
prey. Ecology,77,1975–1988.
Stephens, A. E., Srivastava, D. S., & Myers, J. H. (2013). Strength in
numbers?Effectsofmultiplenaturalenemyspeciesonplantperfor-
mance. Proceedings of the. Royal Society of London B,280,20122756.
Summerville, K. S. (2004). Do smaller forest fragments contain a
greater abundance of lepidopteran crop and forage consumers?
Environmental Entomology,33,234–241.
Thiéry,D.,&Moreau,J.(2005).RelativeperformanceofEuropeangrape-
vine moth (Lobesia botrana)on grapes and other hosts. Oecologia,
143,548–557.
Thomson,L.J.,&Hoffmann,A.A.(2013).Spatialscale ofbenefitsfrom
adjacent woody vegetation on natural enemies within vineyards.
Biological Control,64,57–65.
Tscharntke,T.,Clough,Y.,Wanger,T.C.,Jackson,L.,Motzke,I.,Perfecto,I.,
… Whitbread, A. (2012). Global food security, biodiversity con-
servation and the future of agricultural intensification. Biological
Conservation,151,53–59.
Tscharntke, T., Karp, D. S., Chaplin-Kramer, R., Batáry, P., DeClerck,
F.,Gratton, C., … Zhang, W. (2016). When naturalhabitat fails to
enhance biological pest control – Five hypotheses. Biological
Conservation,204,449–458.
Tscharntke, T., Tylianakis, J. M., Rand,T. A., Didham, R. K., Fahrig, L.,
Batary,P., …Westphal, C. (2012). Landscape moderation of biodi-
versitypatternsandprocesses-eighthypotheses.Biological Reviews,
87,661–685.
Veres,A.,Petit,S.,Conord,C.,&Lavigne,C.(2013).Doeslandscapecom-
positionaffectpestabundanceandtheircontrolbynaturalenemies?
A review. Agriculture, Ecosystems & Environment,166,110–117.
Winqvist,C.,Bengtsson,J.,Aavik,T.,Berendse,F.,Clement,L.W.,Eggers,
S., … Bommarco, R. (2011). Mixed effectsof organic farming and
landscapecomplexityonfarmlandbiodiversityandbiologicalcontrol
potentialacrossEurope.Journal of Applied Ecology,48,570–579.
SUPPORTING INFORMATION
Additional Supporting Information may be found online in the
supportinginformationtabforthisarticle.
How to cite this article:MuneretL,ThiéryD,JoubardB,
RuschA.Deploymentoforganicfarmingatalandscape
scalemaintainslowpestinfestationandhighcrop
productivitylevelsinvineyards.J Appl Ecol. 2017;00:1–10.
https://doi.org/10.1111/1365-2664.13034
