Agriculture Ecosystems and Environment 214 (2015) 46ndash53

Organic farming and host density affect parasitism rates of tortricidmoths in vineyards

A Ruschab L Delbacab L Muneretab D Thieacuteryab

a INRA ISVV UMR 1065 Santeacute et Agroeacutecologie du Vignoble F-33883 Villenave drsquoOrnon FrancebUniversiteacute de Bordeaux UMR 1065 Santeacute et Agroeacutecologie du Vignoble Bordeaux Sciences Agro CS 20032 33882 Villenave drsquoOrnon Cedex France

A R T I C L E I N F O

Article historyReceived 26 January 2015Received in revised form 12 August 2015Accepted 17 August 2015Available online xxx

KeywordsBiological controlPerennial cropEcosystem servicesCrop managementLandscape ecologyFarming systemPestParasitoid

A B S T R A C T

Natural pest control by predators and parasitoids is an important ecosystem service supporting cropproduction It is now well known that the proportion of semi-natural habitats as well as organic farmingenhance abundance and species richness of natural enemies in agroecosystems However few studieshave examined how these environmental variables affect natural pest control services Moreover moststudies have been performed in annual cropping systems and almost nothing is known about the effect oflandscape complexity and organic farming in perennial crops which differ greatly from annual ones interms of disturbance regimes In this study we analyzed how landscape composition and farmingsystems affect abundance of insect pests of grape and their parasitism rates in 79 vineyards insouthwestern France Our results show that farming systems and host density affect biological control oftortricid moths by their parasitoids Surprisingly organic fields had lower parasitism rates compared toconventional ones and this rate was negatively correlated to host density at the field scale We also foundthat moth community composition depended on the proportion of grapevine crop in the landscape in a1 km radius but that pest abundance and parasitism rates did not change with landscape complexity Ourresults suggest that some farming practices that are frequent in organic farming such as organic-certifiedinsecticides copper or sulfur can reduce parasitoid populations and thus limit biological control invineyards Negative density dependence relationship between parasitism rates and host abundancesuggest a dilution effect of the biological control potential at the landscape scale and potentialmechanisms such as variable parasitoid population sizes relatively limited female longevity or fecundityas well as increase in handling time Further research on the effect of organic and conventional farmingpractices are now needed to provide a more mechanistic understanding of how these agriculturalpractices shape ecological processes such as biological control of pests

atilde 2015 Published by Elsevier BV

Contents lists available at ScienceDirect

Agriculture Ecosystems and Environment

journa l homepage wwwe l sev ier com loca te agee

1 Introduction

Agricultural production systems are facing a challenge Risingdemand for agricultural products in yields and quality will increasepressure to further intensify farming systems while there is a needto minimize negative impacts on the environment (Bommarcoet al 2013 Tilman et al 2002) Ecological intensification ofcropping systems thus appears to be a promising alternative tomeet such a challenge by enhancing the services provided bybiodiversity and reducing the negative impacts of agriculture onthe environment (Bommarco et al 2013 Godfray and Garnett

Corresponding author at INRA ndash UMR SAVE ndash Santeacute et Agroeacutecologie duVignoble 71 Avenue Edouard Bourlaux 33883 Villenave drsquoOrnon France Fax +33557122621

E-mail address adrienruschbordeauxinrafr (A Rusch)

httpdxdoiorg101016jagee2015080190167-8809atilde 2015 Published by Elsevier BV

2014) Achieving food security and environmental well-beingtherefore require improved understanding of the factors affectingservice-providing communities and about how to integrate themanagement of ecosystem services into our farming systems(Bommarco et al 2013 Power 2010 Rusch et al 2010)

Natural pest control by predators and parasitoids is animportant ecosystem service supporting crop production (Loseyand Vaughan 2006) It is now well known that this process isaffected by several variables acting at different spatio-temporalscales such as crop management at the field scale or landscapecontext (Rusch et al 2010 Tscharntke et al 2007) A growing bodyof evidence suggests that the proportion of semi-natural habitatsin the landscape strongly influences natural enemy communitiesand trophic interactions in agroecosystems (Bianchi et al 2006Tscharntke et al 2007) and it is now demonstrated that landscapecomplexity enhances abundance and diversity of natural enemies(Chaplin-Kramer et al 2011) that can lead to higher parasitism or

A Rusch et al Agriculture Ecosystems and Environment 214 (2015) 46ndash53 47

predation rates of phytophagous pests (Letourneau et al 2009Rusch et al 2013 Thies et al 2003) This positive effect oflandscape complexity is due to the fact that semi-natural habitatsprovide several key resources for natural enemies such asalternative host and prey nectar overwintering sites or favorablemicroclimatic conditions (Landis et al 2000 Rusch et al 2010Sarthou et al 2014) Thus it is usually assumed that higherproportion of arable land in the landscape will increase pestpressure due to reduced biological control by natural enemies andhigher food resources for pest populations (Meehan et al 2011)However very few studies have considered the effect of landscapecontext on pest populations and pest damage (Chaplin-Krameret al 2011 Rusch et al 2013)

Several studies show that organic farming practices at the fieldscale enhances the abundance and the diversity of natural enemiescompared to conventional farming practices (Bengtsson et al2005 Hole et al 2005 Tuck et al 2014) This effect is usuallyattributed to the use of synthetic pesticides and higher levels ofdisturbance in conventional farming (Bengtsson et al 2005)However each type of farming system encompasses a wide rangeof practices and their relative and combined effects on naturalenemy communities and pest control remain largely unexplored(but see Puech et al 2014) It can be hypothesized that organicfields have higher rates of biological control and higher spatio-temporal stability in the biological control due to higher speciesrichness and functional complementarity between species (Crow-der et al 2010) However a very limited number of studies haveexamined the effect of farming systems on the level of natural pestcontrol and these studies have produced contrasting results(Crowder et al 2010 Lohaus et al 2013 Macfadyen et al2009 Roschewitz et al 2005 Sandhu et al 2010) Moreover it hasbeen recently hypothesized that the effect of organic farming at thelocal scale on biodiversity is modulated by landscape context (ldquotheintermediate landscape complexity hypothesisrdquomdashTscharntke et al2012) According to this hypothesis the benefits of organic farmingat the local scale on biodiversity are smaller in very complexlandscapes (that already support high level of biodiversity) or invery simple landscapes (with a poor species pool) compared tolandscapes of intermediate complexity However this hypothesisremains poorly tested on natural enemy communities andbiological control (but see Rusch et al 2014) There is thereforea need for additional studies in contrasted farming systems tounderstand the relationships between management and function-ing in annual and perennial agroecosystems

In addition to a direct effect of landscape structure on parasitoidpopulations parasitism is also likely to be influenced by hostabundance (Doak 2000) The distribution of parasitism rates inrelation to host density varies between species (Hassell and Waage1984) Some cases reported positive density dependence in whichparasitism rates increased with host density whereas otherreported negative density dependence or density independence(Costamagna et al 2004 Latto and Hassell 1988 Ray andHastings 1996) It has been demonstrated that the relationshipsbetween parasitism rates and host density may vary withparasitoid life-history traits and behavior as well as with thespatial or temporal scales (Doak 2000 Klemola et al 2014 Rolandand Taylor 1997 Teder et al 2000) Several traits of the parasitoidsuch as searching behavior or dispersal abilities could lead todensity dependent parasitism rates Parasitoids use two maincategories of information to localize and parasitize their hostthose related to the resource of their host (plant kairomones orplant habitats characteristics such as plant abundance or shape)and those related to the host itself such as host kairomones (Eschet al 2005 Finch and Collier 2000) Host density is probably themost documented driving factor of parasitoid attraction (Waldeand Murdoch 1988) However most studies document density

dependence mechanisms at the plant scale and almost nothing isknown about density dependence at larger scales (field orlandscape scales for instance) Studying host density at thesescales could lead to different patterns Examining it at the fieldscale could lead to negative density dependence due to dilution ofparasitism capacity because of increase in total handling time oregg depletion while studying host density effects at smaller scales(eg plant) could lead to positive density dependence due toreduced search rate between hosts or natural-enemy aggregation(Rothman and Darling 1991 Walde and Murdoch 1988)Moreover because landscape structure is known to directly affectpest populations as well as their natural enemies it is of majorimportance to disentangle the relative effect of host density andlandscape context on the level of natural pest control Howeverthis remains largely unknown as very few studies have examinedtheir relative and interactive effects (but see Costamagna et al2004)

Four tortricid moths species are usually found in Europeanvineyards and are distributed mainly according to their climaticrequirements the European grapevine moth Lobesia botrana(Denis and Schiffermuumlller) (Lepidoptera Tortricidae) the grapeberry moth Eupoecilia ambiguella (Huumlbner) and the grape tortixArgyrotaenia ljungiana (Thunberg) are polyvoltine while the leaf-rolling tortrix Sparganothis pilleriana (Denis amp Schiffermuumlller) isunivoltine These species are the major grapevine pests in Europeand larvae naturally develop on most grapevine cultivars (Thieacuteryand Moreau 2005 Thieacutery et al 2014) Larvae are polyphagous andcan feed on berries (L botrana and E ambiguella) or on leaves andberries (A ljungiana and S pilleriana) Several alternative hostplants such as ClematisLonicera Ampelopsis or Cormus species areknown to occur in semi-natural habitats such as woodlot orhedgerows located in the surroundings but their occurrence isunknown (Thieacutery 2008) Even if the larvae are polyphagous Vitisvinifera L is their main host in vineyard-dominated areas (Maherand Thieacutery 2006) A wide range of species are known to be naturalenemies of tortricid moths on grape (Sentenac 2011 Thieacutery et al2001) Insect parasitoids classically found in Europe are eggparasitoids (mainly Trichogrammatidae) and larvalpupal para-sitoids (Ichneumonidae Braconidae Chalcididae PteromalidaeEulophidae Elasmidae Tachinidae) The most frequent andefficient species in European vineyards is the solitary larvalendoparasitoid Campoplex capitator (Aubert) (Xueacutereb and Thieacutery2006) This species is known to diapause in its host is specializedon tortricids and has a rather large dispersal range (Thieacutery 2008)Therefore parasitoid populations may disperse from othervineyards as well as from semi-natural habitats in the surroundingenvironment However the different factors affecting the distri-bution of tortricid moths and their natural enemies have mainlybeen studied at the field scale and nothing is known about theeffect of farming systems and landscape context on thesecommunities (Thieacutery and Moreau 2005 Thieacutery et al 2014)

We studied tortricid moths and their parasitoids as a modelsystem to examine the effect of organic and conventional farmingsystems on pest community composition and parasitism ratesalong a landscape complexity gradient in a vineyard region Wefirst hypothesized that moth community composition depends onlandscape context and particularly that pest abundance wouldincrease with the proportion of grapevine crop in the landscapedue to higher resources availability Based on the literature we alsohypothesized that parasitism rates of tortricid moths would behigher in organic fields and that this positive effect of organicfarming is modulated by the landscape context We expected alarger effect of organic farming on natural pest control inlandscapes of intermediate complexity compared to complexlandscapes in which biodiversity and ecosystem services arealready maximized or to very simple landscapes in which

48 A Rusch et al Agriculture Ecosystems and Environment 214 (2015) 46ndash53

biodiversity and ecosystem services are very low (Tscharntke et al2012) Finally we hypothesized a negative density dependencerelationship between parasitism and host density at the field scaledue to dilution effect on parasitism capacity

2 Materials and methods

21 Study site and design

The study was carried out in Bordeaux area in southwesternFrance during three years between 2009 and 2011 in 79 differentvineyards (Table 1) Some vineyards were surveyed during one yearwhereas others were surveyed during more than one year In theBordeaux area (Aquitaine region) organic vineyards representedabout 5 of the vineyards between 2009 and 2011 but thatproportion is increasing rapidly Vine stock density did not differgreatly between the surveyed vineyards (mean 5495 764 plantha) Landscapes in this area vary from extremely simplified whereonly grapevines are cultivated to complex ones characterized by ahigh proportion of semi-natural habitats such as woodland andgrassland Organic and conventional vineyards were selected alonga landscape complexity gradient to examine the relative effects offarming system and landscape on tortricid abundance on grapesand the extent of pest population reduction by their parasitoids(Table 1) The landscape around each vineyard was characterizedby calculating the proportion of semi-natural habitats andagricultural areas within a 1000 m radius using ArcView 10 (ESRI)and data from the Corine Land Cover database (Buumlttner et al2002) For each sampled vineyard two within-field variables wererecorded the use of insecticide (organic-certified or synthetic)during the last two years (qualitative assessment yes or no) andthe type of grass cover management (qualitative assessment novegetation left half-covered or totally covered) These variablescan potentially affect pest and parasitoid populations and allow abetter description of differences between fields in terms of cropmanagement (Danne et al 2010 Nash et al 2010 Thomson andHoffmann 2007)

22 Sampling method

The larvae of the four tortricid species studied build individualnests in their spring generation The proportion of plant attackedby tortricid moths was thus assessed by scoring the number oflarval nests at bloom on 100 grape clusters (for L botrana Eambiguella and A ljungiana) and on 100 leaves (for S pilleriana)One grape cluster and one leaf per plant were selected on100 randomly chosen plants All the collected samples (larvalnests) were brought to the laboratory and were maintained insmall boxes with freshly collected bunches and larvae were fed adlibitum until the end of their development (Moreau et al 2009)Larval populations were checked until pupation upon whichpupae were removed from the flower buds and isolated in glasstubes and stored at 22 C 60 RH and 168 LD Adult moths and

Table 1Characteristics of the sampled vineyards between 2009 and 2011 Proportions of semi-

Year Farming system Number of vi

2009 Organic 5

Conventional 18

2010 Organic 12

Conventional 20

2011 Organic 13

Conventional 30

parasitoids were identified after emerging using the taxonomic keyof Sentenac (2011)

23 Statistical analysis

231 Tortricid moths and parasitoid community compositionTo analyze how farming systems and the proportion of vineyard

affect tortricid moths and parasitoid community composition(relative species abundance) we used PERMANOVA on a Hellingerdistance matrix between sites using the lsquoveganrsquo package in R(function lsquoadonisrsquo) (Oksanen et al 2013) Hellinger distance wasused to reduce the effect of high abundance values in the matrix Ayear effect was also added to examine change in speciescomposition between years PERMANOVA results were calculatedbased on 999 permutations To test for multivariate dispersion weused the lsquobetadisperrsquo function from the lsquoveganrsquo package Tovisualize patterns in species composition we used non-metricmultidimensional (NMDS)

232 Pest abundance and moth parasitism ratesGeneralized linear mixed effect model (GLMM) with quasi-

binomial error distribution was fitted to analyze the effects offarming systems and the proportion of vineyards in the 1 km radiuson the proportion of grape with tortricid larvae (function glmmPQLfrom R package MASS) A quasi-binomial distribution was used toaccount for overdispersion Following this analysis the use ofinsecticide during the two preceding years (yes or no) was added inthe full model in order to explicitly test for the effect of the use ofinsecticides (either synthetic or organic-certified) on pest pres-sure independently of the farming system

GLMM with binomial error distribution was fitted to examinethe effect of host density farming systems and the proportion ofsemi-natural habitats on overall mean parasitism rates of grapeberry moths Following this analysis two within-field variableswere added in the full model (the use of insecticide during theprevious two preceding years (yes or no) and the type of grasscover between rows (no vegetation half-covered and totallycovered)) in order to explicitly test for the effect of these variableson biological control independently of the farming system Inaddition to overall mean parasitism rates of grape berry moths wealso built models for parasitism rates from C capitator (the largelydominant parasitoid species in our communities) Because non-linear interaction between landscape complexity and local farmingsystem could be expected we compared models (GLMMs)including a quadratic or an exponential term of landscapecomplexity Non-linear terms and their interactions with localmanagement were never significant and model fits (assessed usingAIC) were always worst for GLMM including non-linearity termsOnly GLMM without non-linear terms are therefore presented

In all GLMM year and site were included as nested randomeffect to account for repeated measures in the same year and in thesame site during several years and the structure of fixed effectswas simplified by backward elimination of non-significant

natural habitats were calculated with the Corine land cover database

neyards Proportion of semi-natural habitats (range in)

0ndash280ndash100

0ndash1000ndash100

0ndash910ndash95

A Rusch et al Agriculture Ecosystems and Environment 214 (2015) 46ndash53 49

interactions and main effects (p gt 005) In the results the minimumadequate models for each response variable are presented Weused the R packages lsquolme4rsquo (function lsquoglmerrsquo) and lsquoMASSrsquo (functionlsquoglmmPQLrsquo) to build GLMMs Fields where very low larvalpopulations (lt10 individuals) were collected were not consideredwhen analyzing moth and parasitoid community composition aswell as parasitism rates Grape cultivar was not included as anexplanatory variable because vineyards were largely dominated byone cultivar (more than 65 of the vineyards sampled used lsquoMerlotrsquocultivar) and because it has been previously demonstrated thatgrape cultivar does not affect tortricid moths abundance andparasitism rates (Thieacutery et al 2014 Xueacutereb and Thieacutery 2006)

3 Results

The mean proportion of grape stock attacked by tortricid mothsper field (SD) was 162 132 (min 3 max 46 ) in 2009157 1827 (min 1 max 70) in 2010 and 301 293 (min1 max 100) in 2011 A total of 2745 moth larvae were collectedduring the study among which 365 were parasitized The meanparasitism rate per field was 231 22 (min 0 max 100)From the unparasitized larvae we identified the four main tortricidspecies mentioned above These were dominated by L botrana(76) and E ambiguella (23) whereas A ljungiana and Spilleriana were rarely found (lt1 of the unparasitized larvae)Most of the parasitized larvae were parasitized by C capitator ( 91) Other parasitoid species were Phytomyptera nigrina (lt1)Tranosomella praerogator (lt2) Diadegma fenestrale (lt1) Agry-pon anxium (lt1) Charmon sp (lt1) Elasmus sp (lt1)Approximately 5 of parasitoids were not identified due tomortality at the nymphal instar They all had the characteristicsof the Hymenoptera especially those of the Ichneumonideanymphs but no other morphological traits allowed a more preciseidentification

31 Tortricid moths and the parasitoid community composition

Community composition of tortricid moths was not affected byfarming systems but changed along the proportion of vineyard inthe 1 km radius (Table 2) The fit of significant environmentalvariables to moth community NMDS scores indicated that changein community composition mainly came from an increasedabundance of E ambiguella with the proportion of vineyard inthe landscape whereas the abundance of L botrana was notaffected by landscape context (Fig1) Our results did not reveal anydifferences in terms of community composition between the threesuccessive years but indicated changes in composition betweenyears depending on the farming system (significant farmingsystem year interaction Table 2) Our results revealed thatcommunity composition of parasitoids did not change among

Table 2Effects of farming system proportion of vineyard in a 1 km radius and sampling year on toon permutational analysis of covariance Significant effects are shown in bold (P lt 001

Variables Totricid moth

df F

Farming syst 1 008

Proportion of vineyard 1 1251

Year 2 898

Farming syst proportion of vineyard 1 106

Farming syst year 2 705

Proportion of vineyard year 2 227

Farming syst proportion of vineyard year 2 013

Residuals 34

Total 45

farming systems landscape context or between years (Table 2) Infact parasitoid communities were largely dominated by Ccapitator The occurrence of P nigrina should be noticed as anew species in the Bordeaux Area this species expanding itsgeographical distribution from the South

32 Tortricid moths pressure

No differences in the proportion of grapes attacked by mothswere found between organic and conventional vineyards (Table S1Fig 2) or between vineyards that received at least one insecticide(organic-certified or synthetic) during the last two years or not(Fig S1 in the appendices) Similarly the proportions of grapesattacked by tortricid moths were not related to the proportion ofvineyard in the 1 km radius (Table S1 Fig 2) Indeed none of thecandidate explanatory variables was retained in the minimalgeneralized linear mixed-effect model after model simplification

33 Parasitism rates

The minimal generalized linear mixed-effect model explainingthe proportion of parasitized larvae indicated that the overallparasitism rate of tortricid larvae was negatively related to hostabundance and was lower in organic than in conventional fields(Table 3 Fig 3) No effect of the proportion of semi-naturalhabitats the use of insecticides or grass cover management onparasitism rates was found (Fig 4) We found the same results onparasitism rates from the dominant species C capitator Parasitismrate from C capitator was negatively related to host abundance andwas lower in organic than in conventional fields (Table 3 Fig 4)

4 Discussion

In this study we found that farming systems and host density atthe field scale affected parasitism rates of tortricid moths invineyards Surprisingly we found that organic vineyards had lowerparasitism rates compared to conventional ones and that theparasitism rate (mainly by C capitator) was negatively correlated tohost density at the field scale Moth community composition wasaffected by the proportion of vineyard in the landscape butlandscape context did not affect pest pressure nor parasitism rates

Contrary to what was expected parasitism rates of tortricidmoths were higher in conventional vineyards compared to organicones and landscape context did not affect rates of parasitismTaking into account insecticide use (organic-certified or synthetic)and grass cover management irrespectively of farming systemsdid not help to explain differences in parasitism rates suggestingthat this difference is derived from other co-variables related tofarming systems It has been well demonstrated that organicfarming increases species richness and abundance of different taxa

rtricid moth community composition and parasitoid community composition based)

Parasitoid

P-value df F P-value

0741 1 013 0950007 1 001 0500153 2 331 0130305 1 319 0140005 2 068 0800165 2 539 0100831 2 159 010

2839

Fig 1 Non-metric multidimensional scaling ordination of study sites based on similarity in species composition (2-dimensional stress = 0005) with 95 confidence ellipsesThe clustering of sites in NMDS ordination space indicates that parasitoid composition is similar between farming systems (white squares black ellipse = conventional fieldsblack dots red ellipse = organic fields) Symbols represent site values and confidence ellipses are drawn around the group centroid The angle and length of vector loadingsindicate the direction and strength of associations respectively (For interpretation of the references to color in this figure legend the reader is referred to the web version ofthis article)

50 A Rusch et al Agriculture Ecosystems and Environment 214 (2015) 46ndash53

including natural enemies of crop pests (Bengtsson et al 2005Tuck et al 2014) However the mechanisms behind the effect offarming systems on the level of natural pest control remains poorly

Fig 2 Proportion of grapes attacked by tortricid moths in relation to the proportion osignificant effect of the proportion of vineyard as well as the type of farming system were

referred to the web version of this article)

understood Several studies have examined the effect of organicfarming on the rate of biological control by parasitoids in differentagroecosystems (Lohaus et al 2013 Maalouly et al 2013

f vineyards in the 1 km radius and farming systems (organic or conventional) Nofound (For interpretation of the references to color in this figure legend the reader is

Table 3Summary of the generalized linear mixed-effects model relating host density farming system and landscape complexity to overall parasitism rates and parasitism rates fromCampoplex capitator Results are showing the minimum adequate model resulting from manual backward elimination of non-significant interactions and main effects(P gt 005)

Parasitism rates Variable Estim SE z P

Overall parasitism rates Host density 003 0007 497 lt0001Farming system (=organic) 103 043 235 0018

Campoplex capitator Host density 003 0007 500 lt0001Farming system (=organic) 108 043 248 0013

Fig 3 Relationship between overall parasitism rates of tortricid moths and hostabundance for organic (triangle and solid line) and conventional (circle and dashedline) vineyards The effect of host abundance as well as the difference betweenorganic and conventional farming systems were significant (see Table 3 forstatistics)

Fig 4 Relationship between parasitism rates of moths from Campoplex capitatorand host abundance for organic (triangle and solid line) and conventional (circleand dashed line) vineyards The effect of host abundance as well as the differencebetween organic and conventional farming system were significant (see Table 3 forstatistics)

A Rusch et al Agriculture Ecosystems and Environment 214 (2015) 46ndash53 51

Macfadyen et al 2009 2011 Meyling et al 2013 Roschewitz et al2005) Most of these studies did not detect any differences inparasitism rates between organic and conventional farming evenif the structure of food webs was affected by the type of farmingsystem in some cases Only one study found higher parasitism ratesin organic than in conventional orchards (Maalouly et al 2013)Our study is therefore the first one highlighting a negative effect oforganic farming on the rate of biological control by parasitoidsThis pattern may be due to several characteristics of the systemstudied First both organic and conventional vineyards usedinsecticides (organic-certified and synthetic respectively) that canhave strong impacts on natural enemies (Fig S1) In their studyBahlai et al (2010) found that some organic-approved insecticidehave similar or even greater negative impact on natural enemiesand that they were more detrimental to biological control thanwere some synthetic insecticides Secondly even if organicsystems do not use synthetic pesticides it is known that organicvineyards tend to have higher number of farming interventions(eg mechanical weeding tillage insecticides fungicides) thatmight result in a higher disturbance regime compared toconventional fields depending on the species considered (Delbacet al 2012) Such regimes may have limited natural pest control byparasitoids (Jonsson et al 2012) Moreover the main differencebetween organic and conventional vineyards is the higher use ofcopper and sulfur in organic fields as fungicide against variousgrapevine diseases (unpublished data) These products are knownto affect arthropods in general and natural enemies in particular so

that could explain lower parasitism rates in organic vineyards(Nash et al 2010) For instance several studies have shown thenegative effect of sulphur on a wide range of natural enemiesincluding parasitoids predatory mites and spiders (Gent et al2009 Martinson et al 2001 Nash et al 2010) Our findingssuggest the need for a more detailed description of farmingsystems to provide a more mechanistic understanding of the wayfarming systems shape ecological processes such as biologicalcontrol of pests (Puech et al 2014 Rusch et al 2011)

Our results confirmed the initial hypothesis about negativedensity dependence between parasitism rate and host density atthe field scale A range of relationships between host density andparasitism rate have been reported before (Costamagna et al2004 Doak 2000) In a literature survey including 75 studiesWalde and Murdoch (1988) found positive density dependence in23 of the studies negative density dependence in 28 and densityindependence in 49 The spatial scales at which the studies wereperformed and differences in life-history traits of parasitoidspecies are potential explanations for these variable effects Thenegative density dependence observed at the field scale combinedwith the fact that landscape complexity did not affect pestabundance and parasitism rates suggests different mechanismsincluding variable population sizes of the parasitoids colonizing agiven patch relatively limited female longevity andor fecundity(temporally egg-limited) or increase in handling time Indeed suchmechanisms have been found to be responsible for the negativedensity dependence pattern in hostndashparasitoid interactions

52 A Rusch et al Agriculture Ecosystems and Environment 214 (2015) 46ndash53

(Heimpel and Rosenheim 1998 Walde and Murdoch 1988) Ourresults may also suggest relatively low mobility between patchesat the landscape scale or at least low spillover between low-density(where host population is largely exploited) and high-densitypatches (where host population is underexploited) If spilloverbetween patches were important then they would have compen-sated for low parasitism rates in high-density patches and nodensity-dependence would have been observed However moredetailed studies about life-history traits of the main parasitoidspecies C capitator and particularly about the patterns of hostpatch use and factors influencing the behavior and the fitness ofthe parasitoid in the field are necessary to reveal the exactmechanisms involved in the negative density dependence patternfound here

Our results did not support our hypothesis about higher pestpressure in simple landscape due to lower biological control bytheir parasitoids andor higher food availability as pest abundanceand parasitism rates did not vary along the landscape complexitygradient Moreover our findings did not support the intermediatelandscape-complexity hypothesis which suggests a larger effect oforganic farming on ecosystem services in landscapes of interme-diate complexity compared to extremely simple or complexlandscapes as no interacting effect of landscape complexity andfarming system was detected Even if overall pest abundance didnot change with the landscape context we found that communitycomposition of tortricid larvae was affected by the proportion ofvineyard in the 1 km radius with increased abundance of Eambiguella in landscapes supporting a higher density of vineyardsHowever L botrana remained the dominant species in anylandscape context This result suggests that E ambiguellaresponded to an increase in resources availability in the landscape

It has been demonstrated that biological control by parasitoidsis often stronger in more complex landscapes because they rely onsemi-natural habitats to find key resources such as alternativehosts nectar resources and refuge from disturbance (Landis et al2000 Rusch et al 2010) Our results indicate that the solitaryparasitoid C capitator the main parasitoid species found in ourstudy may not strongly depend on semi-natural habitats formaintaining populations at the landscape scale Vineyard land-scapes in this study may support enough nectar resources for theparasitoid In the light of knowledge about the biology of thespecies our findings suggest that C capitator may be well adaptedto intensified vineyard landscapes as it is a polyvoltine speciesrelatively well specialized on its host able to parasitized allgenerations of tortricid moth (up to three in our area) and notstrongly dependent on semi-natural habitats

5 Conclusions

Our study demonstrates that the level of biological control oftortricid moths in vineyards is negatively related to hostabundance and by some organic farming practices Organicfarming is often considered to be of low impact favoringbiodiversity and associated ecosystem services (Tuck et al2014) While this has largely been demonstrated in annualcropping systems our results show that the positive effect oforganic farming on ecosystem services is highly context-depen-dent and might not hold in perennial cropping systems such asgrapevines where differences between organic and conventionalfarming may be less pronounced than in annual cropping systemsFurther research on the effects of organic and conventionalfarming systems on other groups of natural enemies and theirtrophic interactions in this cropping system are thus needed Toprovide a more mechanistic understanding of the way farmingsystems shape ecological processes such as biological control ofpests we advocate for a more detailed description of farming

practices within those broad farming categories (see Puech et al2014)

Acknowledgements

We thank Steve Wratten for helpful comments that helpedimproving earlier versions of the manuscript We are grateful toJean-Marie Brustis for his implication in field surveys We thanklrsquoUE Viticole 1086 INRA and the different grapevine growers forallowing us to access to their fields The research project was partlyfunded by the CIVB (Comiteacute Interprofessionnel du Vin deBordeaux) by an INRA Research project (project Aidy) and by agrant from Ecophyto amp the French National Foundation forResearch on Biodiversity (FRB) (SOLUTION project) This researchis part of the Labex Cote research Cluster

Appendix A Supplementary data

Supplementary data associated with this article can be found inthe online version at httpdxdoiorg101016jagee201508019

References

Buumlttner G Feranec J Jaffrain G 2002 CORINE land cover update 2000 Technicalguidelines Technical report 89 European Environment Agency CopenhagenDenmark

Bahlai CA Xue Y McCreary CM Schaafsma AW Hallett RH 2010 Choosingorganic pesticides over synthetic pesticides may not effectively mitigateenvironmental risk in soybeans PLoS One 5 e11250 doihttpdxdoiorg101371journalpone0011250

Bengtsson J Ahnstroumlm J Weibull A-C 2005 The effects of organic agriculture onbiodiversity and abundance a meta-analysis J Appl Ecol 42 261ndash269 doihttpdxdoiorg101111j1365-2664200501005x

Bianchi FJJA Booij CJH Tscharntke T 2006 Sustainable pest regulation inagricultural landscapes a review on landscape composition biodiversity andnatural pest control Proc R Soc London B Biol Sci 273 1715ndash1727 doihttpdxdoiorg101098rspb20063530

Bommarco R Kleijn D Potts SG 2013 Ecological intensification harnessingecosystem services for food security Trends Ecol Evol 28 230ndash238 doihttpdxdoiorg101016jtree201210012

Chaplin-Kramer R OrsquoRourke ME Blitzer EJ Kremen C 2011 A meta-analysis ofcrop pest and natural enemy response to landscape complexity Ecol Lett 14922ndash932 doihttpdxdoiorg101111j1461-0248201101642x

Costamagna AC Menalled FD Landis DA 2004 Host density influencesparasitism of the armyworm Pseudaletia unipuncta in agricultural landscapesBasic Appl Ecol 5 347ndash355 doihttpdxdoiorg101016jbaae200404009

Crowder DW Northfield TD Strand MR Snyder WE 2010 Organic agriculturepromotes evenness and natural pest control Nature 466 109ndash112 doihttpdxdoiorg101038nature09183

Danne A Thomson LJ Sharley DJ Penfold CM Hoffmann AA 2010 Effects ofnative grass cover crops on beneficial and pest invertebrates in Australianvineyards Environ Entomol 39 970ndash978 doihttpdxdoiorg101603EN09144

Delbac L Thiery D Smits N Roudet J Merot A Wery J Fermaud M 2012 Newindicators for multi pests and diseases assessment in Conventional Organic andin-Transition vineyard systems ESA 2012 12th Congress of the EuropeanSociety for Agronomy 338ndash339

Doak P 2000 Habitat patchiness and the distribution abundance and populationdynamics of an insect herbivore Ecology 81 1842ndash1857

Esch S Klinkhamer PGL Meijden E van der 2005 Do distances among hostpatches and host density affect the distribution of a specialist parasitoidOecologia 146 218ndash226 doihttpdxdoiorg101007s00442-005-0214-1

Finch S Collier RH 2000 Host-plant selection by insects ndash a theory based onappropriateinappropriate landings by pest insects of cruciferous plantsEntomol Exp Appl 96 91ndash102

Gent DH James DG Wright LC Brooks DJ Barbour JD Dreves AJ Fisher GC Walton VM 2009 Effects of powdery mildew fungicide programs ontwospotted spider mite (Acari Tetranychidae) Hop Aphid (HemipteraAphididae) and their natural enemies in hop yards J Econ Entomol 102 274ndash286 doihttpdxdoiorg1016030291020137

Godfray HCJ Garnett T 2014 Food security and sustainable intensificationPhilos Trans R Soc London B Biol Sci 369 20120273 doihttpdxdoiorg101098rstb20120273

Hassell MP Waage JK 1984 Hostndashparasitoid population interactions Annu RevEntomol 29 89ndash114 doihttpdxdoiorg101146annureven29010184000513

Heimpel GE Rosenheim JA 1998 Egg limitation in parasitoids a review of theevidence and a case study Biol Control 11 160ndash168 doihttpdxdoiorg101006bcon19970587

A Rusch et al Agriculture Ecosystems and Environment 214 (2015) 46ndash53 53

Hole DG Perkins AJ Wilson JD Alexander IH Grice PV Evans AD 2005Does organic farming benefit biodiversity Biol Conserv 122 113ndash130 doihttpdxdoiorg101016jbiocon200407018

Jonsson M Buckley HL Case BS Wratten SD Hale RJ Didham RK 2012Agricultural intensification drives landscape-context effects on hostndashparasitoidinteractions in agroecosystems J Appl Ecol 49 706ndash714 doihttpdxdoiorg101111j1365-2664201202130x

Klemola T Andersson T Ruohomaumlki K 2014 Delayed density-dependentparasitism of eggs and pupae as a contributor to the cyclic population dynamicsof the autumnal moth Oecologia 1751211ndash1225 doihttpdxdoiorg101007s00442-014-2984-9

Landis DA Wratten SD Gurr GM 2000 Habitat management to conservenatural enemies of arthropod pests in agriculture Annu Rev Entomol 45 175ndash201 doihttpdxdoiorg101146annurevento451175

Latto J Hassell MP 1988 Generalist predators and the importance of spatialdensity dependence Oecologia 77 375ndash377 doihttpdxdoiorg101007BF00378045

Letourneau DK Jedlicka JA Bothwell SG Moreno CR 2009 Effects of NaturalEnemy Biodiversity on the Suppression of Arthropod Herbivores in TerrestrialEcosystems Annu Rev Ecol Evol Syst 40 573ndash592 doihttpdxdoiorg101146annurevecolsys110308120320

Lohaus K Vidal S Thies C 2013 Farming practices change food web structures incereal aphid-parasitoid-hyperparasitoid communities Oecologia 171 249ndash259doihttpdxdoiorg101007s00442-012-2387-8

Losey JE Vaughan M 2006 The Economic Value of Ecological Services Providedby Insects BioScience 56 311ndash323 doihttpdxdoiorg1016410006-3568(2006)56[311TEVOES]20CO2

Maalouly M Franck P Bouvier J-C Toubon J-F Lavigne C 2013 Codling mothparasitism is affected by semi-natural habitats and agricultural practices atorchard and landscape levels Agric Ecosyst Environ 169 33ndash42 doihttpdxdoiorg101016jagee201302008

Macfadyen S Gibson R Polaszek A Morris RJ Craze PG Planqueacute RSymondson WOC Memmott J 2009 Do differences in food web structurebetween organic and conventional farms affect the ecosystem service of pestcontrol Ecol Lett 12 229ndash238 doihttpdxdoiorg101111j1461-0248200801279x

Macfadyen S Gibson RH Symondson WOC Memmott J 2011 Landscapestructure influences modularity patterns in farm food webs consequences forpest control Ecol Appl 21 516ndash524 doihttpdxdoiorg10189009-21111

Maher N Thieacutery D 2006 Daphne gnidium a possible native host plant of theEuropean grapevine moth Lobesia botrana stimulates its oviposition Is a hostshift relevant Chemoeoclogy 16 135ndash144 doihttpdxdoiorg101007s00049-006-0339-7

Martinson T Williams III L English-Loeb G 2001 Compatibility of chemicaldisease and insect management practices used in New York vineyards withbiological control by Anagrus spp (Hymenoptera Mymaridae) parasitoids ofErythroneura leafhoppers Biol Control 22 227ndash234 doihttpdxdoiorg101006bcon20010975

Meehan TD Werling BP Landis DA Gratton C 2011 Agricultural landscapesimplification and insecticide use in the Midwestern United States Proc NatlAcad Sci 108 11500ndash11505 doihttpdxdoiorg101073pnas1100751108

Meyling NV Navntoft S Philipsen H Thorup-Kristensen K Eilenberg J 2013Natural regulation of Delia radicum in organic cabbage production AgricEcosyst Environ 164 183ndash189 doihttpdxdoiorg101016jagee201209019

Moreau J Richard A Benrey B Thieacutery D 2009 Host plant cultivar of thegrapevine moth Lobesia botrana affects the life history traits of an egg parasitoidBiol Control 50 117ndash122 doihttpdxdoiorg101016jbiocontrol200903017

Nash MA Hoffmann AA Thomson LJ 2010 Identifying signature of chemicalapplications on indigenous and invasive nontarget arthropod communities invineyards Ecol Appl 20 1693ndash1703 doihttpdxdoiorg10189009-10651

Oksanen J Guillaume Blanchet F Kindt R Legendre P Minchin Peter R OrsquoHaraRB Simpson GL Solymos P Stevens MH H Wagner H 2013 VeganCommunity Ecology Package R package version 2 0ndash10 httpCRANR-projectorgpackage=veganhttpCRANR-projectorgpackage=vegan

Power AG 2010 Ecosystem services and agriculture tradeoffs and synergiesPhilos Trans R Soc B Biol Sci 365 2959ndash2971 doihttpdxdoiorg101098rstb20100143

Puech C Baudry J Joannon A Poggi S Aviron S 2014 Organic vs conventionalfarming dichotomy does it make sense for natural enemies Agric EcosystEnviron 194 48ndash57 doihttpdxdoiorg101016jagee201405002

Ray C Hastings A 1996 Density dependence are we searching at the wrongspatial scale J Anim Ecol 65 556ndash566 doihttpdxdoiorg1023075736

Roland J Taylor PD 1997 Insect parasitoid species respond to forest structure atdifferent spatial scales Nature (London) 386 710ndash713

Roschewitz I Huumlcker M Tscharntke T Thies C 2005 The influence of landscapecontext and farming practices on parasitism of cereal aphids Agric EcosystEnviron 108 218ndash227

Rothman LD Darling DC 1991 Spatial density dependence effects of scale hostspatial pattern and parasitoid reproductive strategy Oikos 62 221ndash230 doihttpdxdoiorg1023073545268

Rusch A Valantin-Morison M Sarthou J Roger-Estrade J 2010 Biologicalcontrol of insect pests in agroecosystems effects of crop management farmingsystems and semi-natural habitats at the landscape scale A review Adv Agron109 219ndash259

Rusch A Valantin-Morison M Sarthou J-P Roger-Estrade J 2011 Multi-scaleeffects of landscape complexity and crop management on pollen beetleparasitism rate Landscape Ecol 26 473ndash486 doihttpdxdoiorg101007s10980-011-9573-7

Rusch A Bommarco R Jonsson M Smith HG Ekbom B 2013 Flow and stabilityof natural pest control services depend on complexity and crop rotation at thelandscape scale J Appl Ecol 50 345ndash354 doihttpdxdoiorg1011111365-266412055

Rusch A Birkhofer K Bommarco R Smith HG Ekbom B 2014 Managementintensity at field and landscape levels affects the structure of generalist predatorcommunities Oecologia 175 971ndash983 doihttpdxdoiorg101007s00442-014-2949-z

Sandhu HS Wratten SD Cullen R 2010 The role of supporting ecosystemservices in conventional and organic arable farmland Ecol Complexity EcosystServ Bridging Ecol Econ Social Sci 7 302ndash310 doihttpdxdoiorg101016jecocom201004006

Sarthou J-P Badoz A Vaissiegravere B Chevallier A Rusch A 2014 Local more thanlandscape parameters structure natural enemy communities during theiroverwintering in semi-natural habitats Agric Ecosyst Environ 194 17ndash28 doihttpdxdoiorg101016jagee201404018

Sentenac G 2011 La faune auxiliaire des vignobles de France ed BrocheacuteTeder T Tanhuanpaumlauml M Ruohomaumlki K Kaitaniemi P Henriksson J 2000

Temporal and spatial variation of larval parasitism in non-outbreakingpopulations of a folivorous moth Oecologia 123 516ndash524 doihttpdxdoiorg101007s004420000346

Thieacutery D Moreau J 2005 Relative performance of European grapevine moth(Lobesia botrana) on grapes and other hosts Oecologia 143 548ndash557

Thieacutery D Monceau K Moreau J 2014 Different emergence phenology ofEuropean grapevine moth (Lobesia botrana Lepidoptera Tortricidae) on sixvarieties of grapes Bull Entomol Res 104 277ndash287 doihttpdxdoiorg101017S000748531300031X

Thieacutery D 2008 Les tordeuses nuisibles au vignoble Les ravageurs de la VigneFeacutereacutet Bordeaux France

Thieacutery D Xuersquoreb A Villemant C Sentenac G Delbac L Kuntzman P 2001Larval parasites of vineyards tortricids a brief overview from 3 French vinegrowing areas IOBCWPRS Bull 24 135ndash142

Thies C Steffan-Dewenter I Tscharntke T 2003 Effects of landscape context onherbivory and parasitism at different spatial scales Oikos 101 18ndash25

Thomson LJ Hoffmann AA 2007 Effects of ground cover (straw and compost) onthe abundance of natural enemies and soil macro invertebrates in vineyardsAgric For Entomol 9 173ndash179 doihttpdxdoiorg101111j1461-9563200700322x

Tilman D Cassman KG Matson PA Naylor R Polasky S 2002 Agriculturalsustainability and intensive production practices Nature 418 671ndash677

Tscharntke T Bommarco R Clough Y Crist TO Kleijn D Rand TA Tylianakis JM Nouhuys S van Vidal S 2007 Conservation biological control and enemydiversity on a landscape scale Biol Control 43 294ndash309 doihttpdxdoiorg101016jbiocontrol200708006

Tscharntke T Tylianakis JM Rand TA Didham RK Fahrig L Bataacutery PBengtsson J Clough Y Crist TO Dormann CF Ewers RM Fruumlnd J Holt RD Holzschuh A Klein AM Kleijn D Kremen C Landis DA Laurance WLindenmayer D Scherber C Sodhi N Steffan-Dewenter I Thies C van derPutten WH Westphal C 2012 Landscape moderation of biodiversity patternsand processesmdasheight hypotheses Biol Rev 87 661ndash685 doihttpdxdoiorg101111j1469-185X201100216x

Tuck SL Winqvist C Mota F Ahnstroumlm J Turnbull LA Bengtsson J 2014Land-use intensity and the effects of organic farming on biodiversity ahierarchical meta-analysis J Appl Ecol 51 746ndash755 doihttpdxdoiorg1011111365-266412219

Walde SJ Murdoch WW 1988 Spatial density dependence in parasitoids AnnuRev Entomol 33 441ndash466 doihttpdxdoiorg101146annureven33010188002301

Xueacutereb A Thieacutery D 2006 Does natural larval parasitism of Lobesia botrana(Lepidoptera Tortricidae) vary between years generation density of the hostand vine cultivar Bull Entomol Res 96 105ndash110 doihttpdxdoiorg101079BER2005393