Received: 16 March 2018 | Accepted: 29 September 2018 DOI: 10.1111/1365-2664.13292 RESEARCH ARTICLE Management trade-­offs on ecosystem services in apple orchards across Europe: Direct and indirect effects of organic production Ulrika Samnegård1 | Georgina Alins2 | Virginie Boreux3 | Jordi Bosch4 | Daniel García5 | Anne-Kathrin Happe6 | Alexandra-Maria Klein3 | Marcos Miñarro7 | Karsten Mody6 | Mario Porcel8 | Anselm Rodrigo4 | Laura Roquer-Beni4 | Marco Tasin8 | Peter A. Hambäck1 1 Department of Ecology, Environment and Plant Sciences, Stockholm University, Stockholm, Sweden; 2Institut de Recerca i Tecnologia Agroalimentàries, Lleida, Spain; 3Nature Conservation and Landscape Ecology, University of Freiburg, Freiburg, Germany; 4CREAF, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Spain; 5Unidad Mixta de Investigación en Biodiversidad (CSIC-UO-PA), Universidad de Oviedo, Oviedo, Spain; 6Department of Biology, Technische Universität Darmstadt, Darmstadt, Germany; 7Servicio Regional de Investigación y Desarrollo Agroalimentario, Asturias, Spain and 8Department of Plant Protection Biology, Swedish University of Agricultural Sciences, Alnarp, Sweden Correspondence Peter Hambäck Abstract Email: peter.hamback@su.se 1. Apple is considered the most important fruit crop in temperate areas and profit- Present address able production depends on multiple ecosystem services, including the reduction Ulrika Samnegård, Department of of pest damage and the provision of sufficient pollination levels. Management ap- Biology, Lund University, Lund, Sweden proaches present an inherent trade-off as each affects species differently. Funding information 2. We quantified the direct and indirect effects of management (organic vs. inte- BiodivERsA-FACCE-JPI, Grant/Award grated pest management, IPM) on species richness, ecosystem services, and fruit Number: 2014-74; Svenska Forskningsrådet Formas, Grant/Award Number: 2013-934 production in 85 apple orchards in three European countries. We also quantified and 2014-1784; Ministerio de Economía how habit composition influenced these effects at three spatial scales: within or- y Competitividad, Grant/Award Number: PCIN-2014-145-C02; German Federal chards, adjacent to orchards, and in the surrounding landscape. Ministry of Education and Research, Grant/ 3. Organic management resulted in 48% lower yield than IPM, and also that the varia- Award Number: 01LC1403; Instituto Nacional de Investigación y Tecnología tion between orchards was large with some organic orchards having a higher yield Agraria y Alimentaria than the average yield of IPM orchards. The lower yield in organic orchards resulted Handling Editor: Jennifer Firn directly from management practices, and from higher pest damage in organic or- chards. These negative yield effects were partly offset by indirect positive effects from more natural enemies and higher flower visitation rates in organic orchards. 4. Two factors other than management affected species richness and ecosystem services. Higher cover of flowering plants within and adjacent to the apple trees increased flower visitation rates by pollinating insects and a higher cover of apple orchards in the landscape decreased species richness of beneficial arthropods. 5. The species richness of beneficial arthropods in orchards was uncorrelated with fruit production, suggesting that diversity can be increased without large yield This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2018 The Authors. Journal of Applied Ecology published by John Wiley & Sons Ltd on behalf of British Ecological Society 802 | wileyonlinelibrary.com/journal/jpe  J Appl Ecol. 2019;56:802–811. SAMNEGÅRD et al. Journal of Applied Ecology | 803 loss. At the same time, organic orchards had 38% higher species richness than IPM orchards, an effect that is likely due to differences in pest management. 6. Synthesis and applications. Our results indicate that organic management is more efficient than integrated pest management in developing environmentally friendly apple orchards with higher species richness. We also demonstrate that there is no inherent trade-off between species richness and yield. Development of more en- vironmentally friendly means for pest control, which do not negatively affect pol- lination services, needs to be a priority for sustainable apple production. KEYWORDS apple production, biological control, integrated pest management, natural enemies, organic management, pollination services, species richness, structural equation model 1 | I NTRO D U C TI O N leading to a vulnerable system if managed bees are relied upon to provide the majority of the pollination services (Breeze, Bailey, Fruit consumption is an important part of human nutrition, and Balcombe, & Potts, 2011). the second most important fruit crop globally is apple (FAO, Agricultural intensification affects beneficial arthropods, and 2018). Therefore, the sustainable production of apples is an their delivery of ecosystem services, not only due to local manage- important goal for human food provisioning. In temperate re- ment but also through simplification of the surrounding landscape gions, the by far largest area of fruit production is apple or- (Lichtenberg et al., 2017). The abundance of both natural ene- chards and, similar to other crops, agricultural intensification mies and pollinators is often lower in simplified landscapes, due of these orchards during the last century has increased produc- to lower amounts of alternative resources or fewer overwintering tion through high input of inorganic fertilizers, pesticides, and sites (Shackelford et al., 2013), but there is also often an interac- herbicides (Reganold, Glover, Andrews, & Hinman, 2001). For tion between the local management and structure of surrounding instance, chemical pest control is essential for profitable apple habitats. For instance, it seems that the negative effects of inten- production, as more than 50% of the crop may be lost in or- sive field management on pollinating insects are mainly observed chards with no control (Cross, Fountain, Marko, & Nagy, 2015). in relatively homogeneous landscapes (Rundlöf, Nilsson, & Smith, Intensification in apple orchards, however, leads to increased 2008; Williams & Kremen, 2007). production costs as well as to environmental detriments both In the European Union, subsidies have been available since the within the orchard and in surrounding areas (Reganold et al., late 1980s to promote environmentally friendly farming systems, 2001). These detrimental effects have increased the interest in at both local and landscape scales (Primdahl, Peco, Schramek, developing more environmentally friendly production, through Andersen, & Onate, 2003). These agri-­e nvironmental schemes, either integrated production or organic management, in which which are mainly implemented on a voluntary basis, include “en- the enhancement of ecosystem services from natural enemies vironmentally favourable extensification of farming”, “integrated can partly replace the use of chemical pesticides in suppressing farm management and organic agriculture”, and “preservation of pest populations (Dib, Sauphanor, & Capowiez, 2016; Simon, landscape and historical features such as hedgerows, ditches, Bouvier, Debras, & Sauphanor, 2010). and woods”. Even though subsidies have been in place for some The intensification of agriculture also threatens the deliv- time, their efficiency to promote biodiversity, and how they af- ery of pollination services from the wild pollinator community fect ecosystem services and yield in apple production systems are (Klein, Fornoff, Mupepele, & Boreux, 2018; Potts et al., 2010). For less clear (but see Albert, Franck, Gilles, & Plantegenest, 2017). pollinator-­d ependent crops such as apple, decreased pollination A problem with implementing efficient management strategies services result in lower seed and fruit set and in a lower profitabil- is that ecosystem services are often differently affected by the ity for the farmer (Garratt et al., 2016; Klein et al., 2018; Mallinger same management action (Shackelford et al., 2013). Different & Gratton, 2015). To obtain better pollination, orchard owners responses for diversity-­r elated ecosystem services to the same often use managed pollinators such as honeybees, and in some management action may be expected because species vary in cases bumblebees. However, the efficiency of these managed pol- their life history, but maximizing the total output of ecosystem linators is debated, and is often found to be lower than that of wild services on apple production necessitates that potential trade-­ pollinators (Garratt et al., 2016; Mallinger & Gratton, 2015). The offs arising from management are identified and accounted for availability of managed pollinators may also vary between years (Power, 2010). 804 | Journal of Applied Ecology SAMNEGÅRD et al. One basic trade-­off between ecosystem services and agricul- 2 | M ATE R I A L S A N D M E TH O DS ture emerges when management that aims to increase crop yield by stimulating plant growth (e.g., by adding nutrients and water, or 2.1 | Study regions by removing competing weeds) also indirectly reduce production The study included three important apple-­growing regions; north-­ by affecting the ecosystem services of pest control and pollination eastern Spain (SP), southern Germany (GE), and southernmost (Power, 2010). Trade-­offs also occur in management aimed to af- Sweden (SW) (Figure 1). In Spain, we selected apple orchards lo- fect diversity-­related services or disservices (positive and negative cated in the provinces of Lleida and Girona, Catalonia. In Germany, effects from biodiversity, respectively), when actions to promote we selected apple orchards in the lake Constance region, Baden-­ beneficial arthropods also benefit pest species, or when actions Württemberg. In Sweden, we selected apple orchards on the east to reduce pest species also negatively affect beneficial species and west coasts of the county Skåne. The target apple varieties in (Saunders, Peisley, Rader, & Luck, 2016; Tscharntke et al., 2016). the study orchards were common for each region: Gala and Golden For instance, several studies suggest that flower strips, which Delicious in Spain, Braeburn in Germany, and Aroma and the sub- are commonly planted to benefit pollinators and natural enemies variety Amorosa (but included some Ingrid Marie and Rubinola) in (Lichtenberg et al., 2017; Wratten, Gillespie, Decourtye, Mader, & Sweden. Desneux, 2012), may not only affect the potential for pest control Within each region, we selected 28 (SP and SW) or 30 (GE) or- but also pest densities and crop damage (Tscharntke et al., 2016). chards, half of which were managed organically and the other half Other studies suggest that flower strips to enhance natural ene- were managed according to IPM guidelines (Malavolta & Cross, mies are most efficient when placed inside orchards (Saunders & 2009). One Swedish orchard was excluded before analysis because Luck, 2018), but these strips may then compete with apple trees for it had been abandoned. The orchards were selected along a land-­use nutrients and water (Granatstein & Sánchez, 2009). Similarly, pesti- gradient, using forest cover as a proxy, with approximately half the cides may negatively affect natural enemies and pollinators, leading orchards harbouring agri-­environmental structures (e.g., hedgerows, to reduced biocontrol (Dib et al., 2016; Fountain & Harris, 2015) flower strips, margins with ruderal vegetation) in their close sur- and pollination services (Pisa et al., 2015; Stanley et al., 2015). roundings (up to 20 m from the edge of the trees). IPM orchards were Because apple production is often limited by pest damage and pol- managed with a similar crop protection strategy and with foliar and lination, alternative pest control measures without negative effects mineral fertilizers applied at multiple times along the season. Crop on natural enemies and pollinators are preferable. Natural enemies protection in these orchards involved a range of chemicals for pest, and pollinators are generally promoted by retaining sheltering hab- weed, and disease control, but the specific active ingredient differed itats within or next to the production areas or by providing nectar somewhat between countries and orchards. Among the organic or- and pollen resources in the form of planted or conserved flowering chards, the majority were certified in accordance with European or plants present in alleyways, margins, and hedgerows (Campbell, National legislation (Council Regulation (EC) No 834/2007), which Wilby, Sutton, & Wackers, 2017; Miñarro & Prida, 2013). involves more restrictive crop protection strategies and organic In this study, we examined trade-­offs between production and fertilizers. In these orchards, pest control mainly occurred through ecosystem services, and between ecosystem services and disser- use of natural extracts (neem, pyrethrum), micro-­organisms (Bacillus vices, by comparing integrated pest management (IPM) and organic apple production, as a broad classification of management systems. We evaluated the role of management (organic vs. IPM) in a study design accounting for agri-­environmental structures and landscape composition affecting diversity at three spatial scales: within or- chards, adjacent to orchards, and in the surrounding landscape. The variables include both floral resources for pollinators and overwin- tering sites for all arthropods, estimated through the cover of flow- ering plants and the area of agri-­environmental structures within and close to the orchard, and the amounts of bee-­friendly habitats in a larger area around the orchard, which may increase the species pool for the local orchard. We also included the cover of orchard area around each focal orchard, as a measure of the homogenization of the landscape. The study was performed in 85 apple orchards in three European countries (Spain, Germany, and Sweden) to cover regional variation in apple production. We collected data on flower visitation rates, pollination deficits, natural enemies, pests and fruit production, and used a structural equation model to disentangle the direct and indirect effects of management and environmental vari- F I G U R E 1 Map showing the study areas in Sweden (SW), ables on seed set and fruit production. Germany (GE), and Spain (SP) SAMNEGÅRD et al. Journal of Applied Ecology | 805 thuringiensis), viruses (e.g., granulosis virus vs. codling moth), and treatments. For hand pollination, we used pollen from pollinizer through mating disruption of specific insect pests, while fungal con- trees growing within or adjacent to the orchard. Using these data, trol strategies involve compounds such as sulphur and lime sulphur. we estimated the pollination deficit as (seed set of supplementary Thus, the contrast of IPM versus organic management involves dif- pollinated flowers) − (seed set of open-­pollinated flowers) for fruit- ferences not only in the intensity and type of chemicals for pest con- lets in May–June. A positive value implies a pollination deficit, in- trol but also in the input and availability of nutrients for crop plants dicating insufficient pollination services. As an estimate of apple (expectedly higher in IPM than in organic), due to the use of chemical production, we calculated an index based on the fruit set, propor- fertilizers and chemical weed control (de Ponti, Rijk, & van Ittersum, tion damaged fruits, and mean apple weight calculated for apples 2012). A few uncertified orchards in Sweden and Spain were man- collected on three marked branches on five trees per orchard. The aged as under organic guidelines with no chemical inputs, and these production index equals the weight of undamaged fruit per 100 orchards were considered organic in this study. flowers, and was calculated as (the proportion of undamaged apples at harvest) × (mean weight of harvested apples) × (fruit set). Fruit set is the per cent flowers that produced fruits at harvest from 18 2.2 | Field sampling branches per orchard, the proportion of undamaged fruits equals Within orchards, we quantified natural enemies, pollinators, polli- one minus damage (see next paragraph), and mean weight was cal- nation success, pests/pest damage, and fruit production. Fieldwork culated from up to 18 apples per orchard. was performed during 2015, and data collection was adjusted to We estimated pest densities and damage in two ways represent- the annual cycle of apple production in each region. Pollination ing the main pest problems for orchard owners. First, we estimated was studied during flowering and pest incidence and damage were aphid abundance by counting the proportion of branches infested surveyed at relevant pest phenological stages. Due to climate dif- by aphid colonies, for each aphid species separately on 13–60 trees ferences, the timing of data collection varied between the three per orchard. The main aphid pests in all study orchards and in apple countries. We estimated natural enemy abundance by beating one orchard across Europe are rosy apple aphid (Dysaphis plantaginea apple branch of a representative size on 24 trees per orchard once [Passerini], hereafter RAA) and woolly apple aphid (Eriosoma lan- within 2 weeks after apple flowering, and all collected natural en- igerum [Hausmann]) (Blommers, 1994). RAA was by far the most emies were identified to species or morphospecies. Trees used for abundant species, particularly in Sweden and Spain, and is often beating samples were randomly selected in one 40 m transect per considered as the most damaging aphid, so we only considered this orchard, along a single row perpendicular to the orchard border species. Second, we estimated fruit damage from other pest species (SP and SW) or two rows (GE). Natural enemy abundance was cal- for 24 apples on 37 trees per orchard (888 fruits per orchard), in the culated as the total number of natural enemy individuals collected same transects as the pollination study, at the time of harvest, and per transect, and the richness as the total number of natural enemy used these data to calculate the proportion of damaged apples. This species per transect. In natural enemies, we included spiders, preda- measure reflects the damage of codling moth (Cydia pomonella L.), tory coleopterans (mainly Cantharidae and Coccinellidae), earwigs, sawflies, geometrids, and leaf rollers. The specific pests inflicting the predatory heteropterans (mainly Anthocoridae), predatory dipterans damage differed between countries, with leaf rollers and winter moth (mainly Hybotidae, Empididae, and Dolichopodidae), lacewings, and (Operophtera brumata L.) doing most damage in the Swedish orchards, harvestmen. leaf rollers (Tortricidae), and sawflies (Hoplocampa testudinea Klug) in The visitation rate and richness of apple flower-­visiting wild German orchards and codling moth in Spanish orchards. These esti- pollinators was estimated once per orchard from transect walks mates do not cover damage that cause fruit drop before harvest, but during apple flowering, in one transect close to the orchard bor- such loss would be reflected in the fruit set and thus in the apple der (0–20 m), and one transect in the orchard interior (20–40 m production variable. from border). Each walk lasted 5 min and was repeated three times throughout the day (total 30 min sampling per orchard). Visitation 2.3 | Estimating environmental variables rates were calculated as the number of observed pollinator visits per 1,000 flowers per 5 min. We recorded all pollinators visiting To understand the effect of local conditions, we estimated flowering apple flowers, and collected species for identification in the lab- plant cover and the area covered by agri-­environmental structures oratory. We only included wild bee and syrphid fly species in the (AES) within and in the close surroundings of each orchard. First, we estimate of flower visitation rates as other groups (e.g., beetles) are estimated the cover of flowering plants once per orchard as the per unlikely pollinators of apple (Kendall, 1973; Ramirez & Davenport, cent cover of plants attractive to pollinators (hereafter flower cover) 2013). We pooled the species number of flower visitors and natural near the time of apple flowering. Flowering plants include those spe- enemies to obtain an estimate of the total number of beneficial ar- cies flowering at any time during the year and not only at the time of thropod species per orchard (hereafter, beneficial species richness). the survey, to assess the total amount of resources available for pol- To estimate pollination services, we performed a hand pollination linators. To identify plant species attractive to pollinators, we used experiment on three trees per orchard, where each tree had one the BiolFlor Database (Kühn, Durka, & Klotz, 2004). Flower cover branch dedicated to open and one to supplementary pollination was estimated for each species from six 1 × 1 m2 plots between 806 | Journal of Applied Ecology SAMNEGÅRD et al. apple rows and from six plots outside the apple rows and summed (a) across species. Second, we estimated the total surface cover of AES in m2 within 20 m of the transects. AES include hedgerows (includ- ing edges with old trees and tree rows), forest edges, forests (river forests, tree plantations), fallow lands (including abandoned fields), semi-­natural grasslands (terraced field margins, embankments), and orchard meadows. To understand landscape effects, we estimated the propor- tion of bee-­friendly habitat for each orchard within 1 km from the transect centre. We defined bee-­friendly habitats for each country based on expert knowledge, including shrubland, dry land orchards, and abandoned orchards in Spain, orchard meadows in Germany, and semi-­natural grasslands in Sweden. We estimated the cover of apple orchards as the proportion of surface area covered by this crop within 1 km from the transect, as a proxy for homogeneous land- scape composition and land-­use intensity in our apple production regions. To quantify landscape characteristics, we used official digi- tal maps for Spain and Germany (Carreras & Diego, 2009; LGL 2016; (b) SIOSE 2015), spatially explicit data on land use from the Swedish Board of Agriculture (Integrated Administrative Control System, IACS) and Geographic Information Systems and Remote Sensing software ArcView 10.3.1 and MiraMon. 2.4 | Statistical analyses To assess the direct and indirect effects of management, local or- chard conditions, adjacent site conditions, and landscape composition across orchards, we developed a structural equation model (SEM) with fruit production as the endpoint variable. As intermediary variables, we used the total species richness of beneficial arthropods (flower visitors and natural enemies), natural enemy abundance, flower visi- tation rate by wild pollinators, RAA abundance, pest damage at har- vest, and pollination deficit. To build the SEM, we combined seven mixed effects models (lme in the r package nlme) in a piecewise SEM (Lefcheck, 2016), with country as random effect. To reduce the num- F I G U R E 2 (a) Initial and (b) final structural equation model (SEM) ber of variables, we first evaluated each individual lme and removed showing significant direct and indirect paths from management, non-­significant variables describing agri-­environmental or landscape orchard landscape cover and flower cover. The landscape variables composition. Following this, we evaluated each lme by plotting stand- tested in the initial SEM were flower cover, AES cover, orchard ardized residual against fitted values and predictor variables. For pest cover, and cover of bee-­friendly habitats. Arrow thickness in the damage, residual plots indicated heteroscedasticity between manage- final SEM is proportional to the standardized path coefficients (figures next to the paths). The colour of the path indicates the sign ment and between countries. We therefore modelled variance in this of the effect (red = negative, black = positive). The sign connected submodel using the VarIdent option. For apple production and flower to management type refer to organic management relative to IPM. visitation rate, residual plots indicated a loglinear relationship with pre- The model includes correlated errors between natural enemy dictive variables and these variables were log10-­transformed before abundance and richness of beneficial insects (p < 0.0001), and inclusion in the final model. between (rosy apple aphid) RAA abundance and total damage at harvest (p < 0.0001) but these arrows are omitted in the figure We assessed the initial SEM (Figure 2a) by the D-­separation test to detect missing paths and tested the overall model with Fisher’s C statistics. We added significant missing paths and removed non-­ To assess the generality of the model across countries, we ran the significant paths until the AIC was no longer reduced. We accounted final SEM for each country separately as a post hoc comparison. This for two correlated errors; between species richness and natural step should be viewed cautiously as the model is applied on smaller enemy abundance, and between RAA abundance and total damage datasets, but it serves the purpose of indicating if patterns in the at harvest. When presenting the final SEM, we compared the rela- SEM are mainly caused by patterns in one country. In this compar- tive importance of pathways using standardized path coefficients. ison, we present unstandardized parameter values because these SAMNEGÅRD et al. Journal of Applied Ecology | 807 provide a better comparison between countries. To assess the rela- were non-­significant and were excluded already in the initial model. tionship between apple production and species richness, we related The only effects from the agri-­environmental or landscape struc- these variables following the removal of partial effects from other tures that were retained in the final SEM were positive effects of variables in the lme-­models using the remef command (Hohenstein flower cover on wild pollinator visitation rates and negative ef- & Kliegl, 2015). fects of orchard cover on species richness of beneficial arthropods (Figure 2b). When comparing parameter values between countries and 3 | R E S U LT S with the final SEM (Table 1), differences were relatively small. In three cases, parameter values for the three countries deviated When analysing the combined direct and indirect effects of man- based on the difference of parameter values and the magnitude agement on fruit production, we found that organic orchards on av- of the SE. First, estimated parameter values for the relation- erage had a 48% lower fruit production compared to IPM orchards ship between natural enemy abundance and fruit production (F1, 76 = 20.9, p < 0.0001) and this effect size did not vary between was lower for Sweden and did not overlap with the estimates countries (F2, 76 = 2.1, p > 0.13). However, the variation for each for other countries. Second, estimated values for the relation- category was large and the production of the most productive or- ship between management and fruit damage were higher for ganic orchards exceeded the mean of IPM orchards (Figure 3). The Sweden and Spain compared to Germany. Finally, estimated pa- initial SEM showed a good fit (Fisher’s C = 69.4, df = 60, p = 0.18), rameter values for the relationship between management and but the D-­separation test indicated a missing direct path from the aphid abundance (mainly RAA) were lower for Germany than for natural enemy abundance to apple production (for data used in other countries. SEM see Samnegård et al., 2019). Adding this path increased the fit When assessing the relationship between fruit production and of the SEM (Fisher’s C = 47.9, df = 56, p = 0.54), and did not change beneficial arthropod species richness, while partialling out the effect the model otherwise (Figure 2b). In the final SEM, management had of orchard cover in the surrounding landscape, we found no relation- a strong direct effect, and several indirect effects, on apple pro- ship between fruit production and species richness (Figure 4). This duction with lower production for organic orchards. Both natural pattern was true for both organic and IPM orchards, but there was enemy abundance and flower visitation rates were higher in or- an effect of management where organic orchards had on average ganic orchards, creating indirect positive effects from organic man- 38% more species for the same production of apples (13.0 vs. 9.4 agement on apple production (Figure 2b). Fruit damage at harvest species). was higher in organic orchards, creating an indirect negative effect from organic management on apple production (Figure 2b). It is also notable that effects from the area of AES and bee-­friendly habitats 4 | D I S CU S S I O N Management differences between organic and IPM in apple or- 16 Mean Mean organic IPM chards evidently had strong effects on fruit production, pest dam- = 7.7 Organic IPM = 14.8 age, beneficial arthropod species richness, and diversity-­r elated 14 ecosystem services. On average, fruit production was 48% lower 12 in organic orchards, which is a larger difference than between organic and conventional management in other crops (Seufert, 10 Ramankutty, & Foley, 2012). Importantly, the overall effect on No of orchards fruit production was due to a combination of direct management 8 effects and indirect effects due to higher pest damage in organic orchards. Pest control strategies in organic orchards are typically 6 less effective than in orchards that use synthetic pesticides, and the commercial output of apple production is sensitive to dam- 4 age, as damaged fruits cannot be sold as high-­q uality apples for direct consumption on the market. The unidentified direct effects 2 may be related to differences in fertilizer and water/irrigation use 0 (Berry et al., 2002; Klein, Hendrix, Clough, Scofield, & Kremen, 5 10 15 20 25 30 35 40 45 50 2015), as well as fungal disease control and weed management, Apple production index (×100) which were not accounted for in our study. While pest damages were lower in IPM orchards, organic orchards were more strongly F I G U R E 3 Distribution of the fruit production index for organic benefitting from diversity-­r elated ecosystem services, as these orchards and IPM orchards. For illustrative purposes, the index is corrected for differences between countries by multiplying the orchards had both a higher abundance of natural enemies and a value of each orchard by the ratio of the overall and country means higher flower visitation rate from wild pollinators, increasing fruit 808 | Journal of Applied Ecology SAMNEGÅRD et al. TA B L E 1 Unstandardized path coefficients (mean ± SE) of the final model using the whole dataset, with country as random effect, and the three datasets separately. For management, coefficients indicate differences of organic management to IPM Response Predictor Total SP GE SW Fruit production Management −0.19 ± 0.07 (p < 0.005) −0.064 ± 0.11 −0.23 ± 0.14 −0.12 ± 0.15 – Fruit damage −1.72 ± 0.40 (p < 0.004) −1.83 ± 0.51 −1.84 ± 3.12 −2.87 ± 0.95 – NE abundance 0.0062 ± 0.0022 (p < 0.007) 0.0082 ± 0.0050 0.0064 ± 0.0059 0.0009 ± 0.0036 – Flower visitation 0.057 ± 0.028 (p < 0.05) 0.046 ± 0.063 0.004 ± 0.078 0.096 ± 0.075 Pollination deficit Flower visitation −0.77 ± 0.31 (p < 0.03) −1.49 ± 0.45 −0.33 ± 0.65 −0.07 ± 0.51 Flower visitation Flower cover 0.0075 ± 0.0034 (p < 0.03) 0.013 ± 0.005 0.0057 ± 0.0058 −0.0028 ± 0.009 – Management 0.33 ± 0.17 (p < 0.05) 0.32 ± 0.25 0.048 ± 0.33 0.69 ± 0.26 Fruit damage Management 0.034 ± 0.009 (p < 0.0003) 0.11 ± 0.04 0.025 ± 0.008 0.12 ± 0.02 RAA abundance Management 0.26 ± 0.05 (p < 0.0001) 0.36 ± 0.11 −0.011 ± 0.009 0.46 ± 0.10 Richness of Management 4.10 ± 0.78 (p < 0.0001) 5.36 ± 1.32 3.51 ± 0.97 3.18 ± 1.73 beneficials – Orchard cover −0.046 ± 0.019 (p < 0.02) −0.023 ± 0.021 −0.12 ± 0.03 −0.037 ± 0.070 NE abundance Management 6.49 ± 2.46 (p < 0.02) 8.86 ± 3.64 7.00 ± 4.67 3.48 ± 4.35 pattern was similar in both organic and IPM orchards. If anything, there was a close to significant positive relationship between apple production and species richness for IPM orchards. Clough et al. (2011) similarly found no trade-­off between crop yield and biodiversity in smallholder cacao production systems, suggest- ing that productivity costs related to the maintenance of a high biodiversity may be small for some systems. On the other hand, organic orchards in our study had on average 38% more benefi- cial species for similar levels of apple production, confirming the previous conclusion that organic management supports a higher local species richness of arthropods (Rusch, Birkhofer, Bommarco, Smith, & Ekbom, 2014). We can only speculate on the main causes of this difference, but it seems plausible that it is mainly due to differences in pest control methods that affect survival of non-­ F I G U R E 4 Partial residuals, prediction lines, and confidence target organisms (cf., Lefebvre et al., 2017; Park, Blitzer, Gibbs, bands between species richness of beneficial arthropods and the Losey, & Danforth, 2015) or differences in weed management ef- fruit production index (log10-­transformed), for organic (●) and IPM fects on species diversity within the orchards (Gurr, Wratten, & (⚪) orchards Luna, 2003). The effect of orchard management on arthropod richness often production. While these indirect positive effects were not strong interacts with habitat composition at the local and landscape levels, enough to overcome the negative effects of organic management where intensification and homogenization at the landscape level on fruit production, the pattern suggests that methods to increase result in decreased arthropod richness in otherwise species-­rich natural enemy abundance and wild pollinator visitation have the habitats (Landis, 2017). The use of various AES for conservation potential to reduce the yield gap between organic and IPM or- has long been promoted in the European Union (Primdahl et al., chards. This conclusion is supported by the fact that some organic 2003), but the effectiveness of these measures has been ques- orchards had a fruit production that was well above the mean tioned (Batáry, Dicks, Kleijn, & Sutherland, 2015; Tscharntke et al., production of IPM orchards (Figure 3). 2016). In our study, we found that a higher cover of apple orchards A concern for agricultural systems in general has been that in the surrounding landscape reduced species richness of benefi- increased production often causes a reduction in biodiversity, and cial arthropods within the orchard. A relatively uniform landscape that efforts to reduce these negative effects entail a cost in terms with apple orchards may be less favourable for biodiversity than of reduced production (Clough et al., 2011; Gabriel, Sait, Kunin, & a more heterogeneous landscape, because most apple orchards Benton, 2013). Our study does not support this concern for apple in any landscape are managed according to IPM (see also Joshi, orchards. Species richness of beneficial arthropods and apple pro- Otieno, Rajotte, Fleischer, & Biddinger, 2016; Marini, Quaranta, duction in our study were largely uncorrelated (Figure 4), and this Fontana, Biesmeijer, & Bommarco, 2012). On the other hand, we SAMNEGÅRD et al. Journal of Applied Ecology | 809 did not find a direct effect of AES surface on either species rich- Sweden. It is possible that the present natural enemies are less ness of beneficial arthropods, natural enemy abundance, and fruit able to affect winter moth outbreaks. Second, there was a weaker production. However, it is premature to argue that AES are not use- connection between management and pest damage and aphid ful in apple orchards since only a few orchards in our study had abundance in Germany, where aphid control was equally strong actively established these structures. Our measures mainly reflect in both organic and IPM. Aphid densities during the sampling year the natural occurrence of these habitat types in the surroundings of may have been low in Germany for other reasons, reducing the the orchards, and a more targeted establishment of AES may result effect of management. in greater benefits to biodiversity and related ecosystem services. In conclusion, our study shows differences in the delivery of eco- Nevertheless, we found a clear positive direct effect of flower cover system services between organic and IPM apple orchards, where both on pollinator visitation rates of apple flowers (supporting Campbell natural enemy abundance (measuring biocontrol services) and flower et al., 2017), which resulted in reduced pollination deficit (measured visitation rate (measuring pollination services) were higher in organic through seed set) and increased fruit production, suggesting that orchards. Moreover, pollination services were positively affected by targeted establishment of flower strips may have positive effects the flower cover surrounding the orchard. Nevertheless, the average on apple pollination. IPM orchard reached a higher final apple production even though the When examining the role of natural enemies, we found that higher variation between orchards was high and the organic orchard with natural enemy abundance was related to higher fruit production, but the highest production was producing well above the average IPM this effect was not due to a negative relationship between natural orchard. The main reason for the differences in production does not enemy abundance and either aphid abundance or apple pest damage seem to be related to the observed differences in ecosystem services at harvest. This finding suggests that the natural enemies provide some as there was a strong direct (and unexplained) effect of management biocontrol that was not captured by our pest sampling. The community on apple production. Yet our study also suggests that there is scope for of apple pests shows large differences between the different countries, increasing the diversity of beneficial arthropods without reducing pro- and we therefore had to use relatively coarse measures of damage. duction. If differences in species richness between organic and IPM It is possible that the natural enemies found in this study mainly reg- are due mainly to pest control strategies then this pattern would sup- ulated earlier pest insects and that this effect is not reflected in our port a continued focus on developing targeted pest control methods measure of fruit damage. It is also evident that the group of natural that are also environmentally friendly. enemies is heterogeneous, including spiders, coleopterans, dipterans, neuropterans, heteropterans, earwigs, and harvestmen. Different nat- AC K N OW L E D G E M E N T S ural enemies have different diets. Some groups are known to feed on and reduce apple pests (Cross et al., 2015), while the feeding habitats We thank all participating apple growers for allowing us to include their and effects on pest species are less understood for other groups. In orchard in the project, KOB Bavendorf, FÖKO e.V. and Äppelriket in addition to pest species, also natural enemies varied in abundance be- Kivik, ADV Ecològica de Ponent, and ACTEL, and all field assistants. The tween countries, with a higher abundance of dipterans in Sweden and research was part of the EcoFruit project funded through the 2013‐2014 a higher abundance of heteropterans and earwigs in Germany (A. K. BiodivERsA/FACCE‐JPI joint call (agreement# BiodivERsA-­FACCE 2014-­ Happe, N. Blüthgen, V. Boreux, J. Bosch, D. García, P. A. Hambäck, A. M. 74), with the national funders Swedish Research Council Formas (grant# Klein, M. Miñarro, A. Rodrigo, L. Roquer-Beni, U. Samnegård, G. Alins, 2014-­1784), German Federal Ministry of Education and Research M. Porcel, R. Martínez Sastre, M. Tasin and K. Mody, unpublished data). (PT-­DLR/BMBF) (grant# 01LC1403), and the Spanish Ministerio de It is also evident that our focus on arthropod natural enemies ignore Economía y Competitividad (MINECO) (project# PCIN-­2014-­145-­C02). birds, which are known to reduce both caterpillar and aphid damage in The research was also funded by Formas grant# 2013-­934 (to MT) apple (García, Miñarro, & Martínez-­Sastre, 2018; Mols & Visser, 2002). and project RTA2013-­00039-­C03-­00 funded by Instituto Nacional de Regional differences in management, landscape context, and Investigación y Tecnología Agraria y Alimentaria (to GA). The use of IACS in the biota on apple trees may affect the effect of organic man- in Sweden was developed within the projects SAPES and MULTAGRI and agement versus IPM. For instance, Kehinde, Wehrden, Samways, was adapted to our study by M. Stjernman and P. Olsson. Klein, and Brittain (2018) found that the bee abundance in vine- yards was positively affected by organic management in Italy but AU T H O R S ’ C O N T R I B U T I O N S not in South Africa, with potential effects on pollination. In our study, we found surprisingly strong regional similarities when A.-M.K., A.-K.H., J.B., K.M., L.R.-B., P.A.H., U.S., and V.B. conceived comparing organic management and IPM. The SEM coefficients the ideas for the paper and designed the studies; U.S., G.A., V.B., were mostly on the same order with a few exceptions. First, there A.-K.H., M.P., A.R., D.G., M.M., M.T., and L.R.-B. collected the was a weaker connection between the natural enemy density and data; U.S. and P.A.H. analysed the data and led the writing of the fruit damage in Sweden, which may be due to differences in the manuscript. All authors contributed to the development of ideas and pest community where winter moth was a dominant pest only in drafts and gave final approval for publication. 810 | Journal of Applied Ecology SAMNEGÅRD et al. DATA ACC E S S I B I L I T Y FAO (2018). Food and Agriculture Organization of the United Nations, Retrieved from http://www.fao.org/statistics (Accessed 24 February All data for the SEM analysis and the statistical codes are avail- 2018). able via the Dryad Digital Repository https://doi.org/10.5061/ Fountain, M. T., & Harris, A. L. (2015). Non-­t arget consequences of in- secticides used in apple and pear orchards on Forficula auricularia L. dryad.3pc3j00 (Samnegård et al., 2019). (Dermaptera: Forficulidae). Biological Control, 91, 27–33. https://doi. org/10.1016/j.biocontrol.2015.07.007 Gabriel, D., Sait, S. M., Kunin, W. E., & Benton, T. G. (2013). Food pro- ORCID duction vs. biodiversity: Comparing organic and conventional agriculture. Journal of Applied Ecology, 50, 355–364. https://doi. Daniel García http://orcid.org/0000-0002-7334-7836 org/10.1111/1365-2664.12035 García, D., Miñarro, M., & Martínez-Sastre, R. (2018). Birds as suppliers Marco Tasin http://orcid.org/0000-0001-7379-4954 of pest control in cider apple orchards: Avian biodiversity drivers and Peter A. Hambäck http://orcid.org/0000-0001-6362-6199 insectivory effect. Agriculture Ecosystems & Environment, 254, 233– 243. https://doi.org/10.1016/j.agee.2017.11.034 Garratt, M. P. D., Breeze, T. D., Boreux, V., Fountain, M. T., McKerchar, M., Webber, S. M., … Potts, S. G. (2016). Apple pollination: Demand de- REFERENCES pends on variety and supply depends on pollinator identity. PLoS ONE, 11, e0153889. https://doi.org/10.1371/journal.pone.0153889 Albert, L., Franck, P., Gilles, Y., & Plantegenest, M. (2017). Impact of agro- Granatstein, D., & Sánchez, E. (2009). Research knowledge and ecological infrastructures on the dynamics of Dysaphis plantaginea needs for orchard floor management in organic tree fruit sys- (Hemiptera: Aphididae) and its natural enemies in apple orchards tems. International Journal of Fruit Science, 9, 257–281. https://doi. in Northwestern France. Environmental Entomology, 46, 528–537. org/10.1080/15538360903245212 https://doi.org/10.1093/ee/nvx054 Gurr, G. M., Wratten, S. D., & Luna, J. M. (2003). Multi-­function agricultural Batáry, P., Dicks, L. V., Kleijn, D., & Sutherland, W. J. (2015). The role biodiversity: Pest management and other benefits. Basic and Applied of agri-­environment schemes in conservation and environmen- Ecology, 4, 107–116. https://doi.org/10.1078/1439-1791-00122 tal management. Conservation Biology, 29, 1006–1016. https://doi. Hohenstein, S., & Kliegl, R. (2015). Remove partial effects, R package org/10.1111/cobi.12536 version 1.0.6.9000. Berry, P. M., Sylvester-Bradley, R., Philipps, L., Hatch, D. J., Cuttle, Joshi, N. K., Otieno, M., Rajotte, E. G., Fleischer, S. J., & Biddinger, D. J. S. P., Rayns, F. W., & Gosling, P. (2002). Is the productivity of or- (2016). Proximity to woodland and landscape structure drives polli- ganic farms restricted by the supply of available nitrogen? Soil nator visitation in apple orchard ecosystem. Frontiers in ecology and Use and Management, 18, 248–255. https://doi.org/10.1079/ evolution, 4, 38. SUM2002129 Kehinde, T., Wehrden, H., Samways, M., Klein, A. M., & Brittain, C. (2018). Blommers, L. H. (1994). Integrated pest management in European apple Organic farming promotes bee abundance in vineyards in Italy but orchards. Annual Review of Entomology, 39, 213–241. https://doi. not in South Africa. Journal of Insect Conservation, 22, 61–67. https:// org/10.1146/annurev.en.39.010194.001241 doi.org/10.1007/s10841-017-0038-4 Breeze, T. D., Bailey, A. P., Balcombe, K. G., & Potts, S. G. (2011). Kendall, D. A. (1973). Viability and compatibility of pollen on insects vis- Pollination services in the UK: How important are honeybees? iting apple blossom. Journal of Applied Ecology, 10, 847–853. https:// Agriculture Ecosystems & Environment, 142, 137–143. https://doi. doi.org/10.2307/2401873 org/10.1016/j.agee.2011.03.020 Klein, A. M., Fornoff, F., Mupepele, A. C., & Boreux, V. (2018). Relevance Campbell, A. J., Wilby, A., Sutton, P., & Wackers, F. L. (2017). Do sown of wild and managed bees for human well-­being. Current Opinion in flower strips boost wild pollinator abundance and pollination ser- Insect Science, 26, 82–88. https://doi.org/10.1016/j.cois.2018.02.011 vices in a spring-­flowering crop? A case study from UK cider apple Klein, A. M., Hendrix, S. D., Clough, Y., Scofield, A., & Kremen, C. (2015). orchards. Agriculture Ecosystems & Environment, 239, 20–29. https:// Interacting effects of pollination, water and nutrients on fruit tree doi.org/10.1016/j.agee.2017.01.005 performance. Plant Biology, 17, 201–208. https://doi.org/10.1111/ Carreras, J., & Diego, F. (2009). Catalan habitats cartography, 1:50,000. plb.12180 Barcelona, Spain: Departament de Medi Ambient i Habitatge, Kühn, I., Durka, W., & Klotz, S. (2004). BiolFlor – A new plant-­trait data- Generalitat de Catalunya. base as a tool for plant invasion ecology. Diversity and Distributions, Clough, Y., Barkmann, J., Juhrbandt, J., Kessler, M., Wanger, T. C., 10, 363–365. Anshary, A., & Tscharntke, T. (2011). Combining high biodiversity Landis, D. A. (2017). Designing agricultural landscapes for biodiversity-­ with high yields in tropical agroforests. Proceedings of the National based ecosystem services. Basic and Applied Ecology, 18, 1–12. Academy of Sciences of the United States of America, 108, 8311–8316. https://doi.org/10.1016/j.baae.2016.07.005 https://doi.org/10.1073/pnas.1016799108 Lefcheck, J. S. (2016). Piecewise structural equation modelling. R pack- Cross, J., Fountain, M., Marko, V., & Nagy, C. (2015). Arthropod age version 1.2.1. ecosystem services in apple orchards and their economic bene- Lefebvre, M., Franck, P., Olivares, J., Ricard, J. M., Mandrin, J. F., & fits. Ecological Entomology, 40, 82–96. https://doi.org/10.1111/ Lavigne, C. (2017). Spider predation on rosy apple aphid in conven- een.12234 tional, organic and insecticide-­free orchards and its impact on aphid de Ponti, T., Rijk, B., & van Ittersum, M. K. (2012). The crop yield gap populations. Biological Control, 104, 57–65. https://doi.org/10.1016/j. between organic and conventional agriculture. Agricultural Systems, biocontrol.2016.10.009 108, 1–9. https://doi.org/10.1016/j.agsy.2011.12.004 LGL. (2016). ATKIS Digitales Landschaftsmodell, Baden-Württemberg, Dib, H., Sauphanor, B., & Capowiez, Y. (2016). Effect of management Basis-DLM Version 6.0. Landesamt für Geoinformation und strategies on arthropod communities in the colonies of rosy apple Landentwicklung. aphid, Dysaphis plantaginea Passerini (Hemiptera: Aphididae) in Lichtenberg, E. M., Kennedy, C. M., Kremen, C., Batary, P., Berendse, F., south-­e astern France. Agriculture Ecosystems & Environment, 216, Bommarco, R., … Crowder, D. W. (2017). A global synthesis of the 203–206. https://doi.org/10.1016/j.agee.2015.10.003 effects of diversified farming systems on arthropod diversity within SAMNEGÅRD et al. Journal of Applied Ecology | 811 fields and across agricultural landscapes. Global Change Biology, 23, Samnegård, U., Alins, G., Boreux, V., Bosch, J., García, D., Happe, 4946–4957. https://doi.org/10.1111/gcb.13714 A.-K., … Hambäck, P. A. (2019). Data from: Management trade-­offs Malavolta, C., & Cross, J.V. (2009). Guidelines for integrated production on ecosystem services in apple orchards across Europe: Direct and of pome fruits in Europe, IOBC Technical Guideline III (Vol. 47, 4th indirect effects of organic production. Dryad Digital Repository, ed., pp. 1–12). IOPC/wprs Bulletin. https://doi.org/10.5061/dryad.3pc3j00 Mallinger, R. E., & Gratton, C. (2015). Species richness of wild bees, but Saunders, M. E., & Luck, G. W. (2018). Interaction effects between local not the use of managed honeybees, increases fruit set of a pollinator-­ flower richness and distance to natural woodland on pest and bene- dependent crop. Journal of Applied Ecology, 52, 323–330. https://doi. ficial insects in apple orchards. Agricultural and Forest Entomology, 20, org/10.1111/1365-2664.12377 279–287. https://doi.org/10.1111/afe.12258 Marini, L., Quaranta, M., Fontana, P., Biesmeijer, J. C., & Bommarco, R. Saunders, M. E., Peisley, R. K., Rader, R., & Luck, G. W. (2016). Pollinators, (2012). Landscape context and elevation affect pollinator communi- pests, and predators: Recognizing ecological trade-­offs in agroecosys- ties in intensive apple orchards. Basic and Applied Ecology, 13, 681– tems. Ambio, 45, 4–14. https://doi.org/10.1007/s13280-015-0696-y 689. https://doi.org/10.1016/j.baae.2012.09.003 Seufert, V., Ramankutty, N., & Foley, J. A. (2012). Comparing the yields Miñarro, M., & Prida, E. (2013). Hedgerows surrounding organic apple of organic and conventional agriculture. Nature, 485, 229–U113. orchards in north-­west Spain: Potential to conserve beneficial in- https://doi.org/10.1038/nature11069 sects. Agricultural and Forest Entomology, 15, 382–390. https://doi. Shackelford, G., Steward, P. R., Benton, T. G., Kunin, W. E., Potts, S. G., org/10.1111/afe.12025 Biesmeijer, J. C., & Sait, S. M. (2013). Comparison of pollinators and Mols, C. M. M., & Visser, M. E. (2002). Great tits can reduce caterpillar natural enemies: A meta-­analysis of landscape and local effects on damage in apple orchards. Journal of Applied Ecology, 39, 888–899. abundance and richness in crops. Biological Reviews, 88, 1002–1021. https://doi.org/10.1046/j.1365-2664.2002.00761.x https://doi.org/10.1111/brv.12040 Park, M. G., Blitzer, E. J., Gibbs, J., Losey, J. E., & Danforth, B. N. (2015). Simon, S., Bouvier, J. C., Debras, J. F., & Sauphanor, B. (2010). Biodiversity Negative effects of pesticides on wild bee communities can be buff- and pest management in orchard systems. A review. Agronomy for ered by landscape context. Proceedings of the Royal Society B-­Biological Sustainable Development, 30, 139–152. https://doi.org/10.1051/ Sciences, 282, 20150299. https://doi.org/10.1098/rspb.2015.0299 agro/2009013 Pisa, L. W., Amaral-Rogers, V., Belzunces, L. P., Bonmatin, J. M., Downs, SIOSE. (2015). Spanish Habitats Cartography, 1:25,000. Equipo Técnico Nacional C. A., Goulson, D., & Wiemers, M. (2015). Effects of neonicot- SIOSE, Spanish Land Use and Land Cover Information System. Ministerio de inoids and fipronil on non-­t arget invertebrates. Environmental Fomento. Centro Nacional de Información Geográfica, Madrid. Science and Pollution Research, 22, 68–102. https://doi.org/10.1007/ Stanley, D. A., Garratt, M. P. D., Wickens, J. B., Wickens, V. J., Potts, S. G., s11356-014-3471-x & Raine, N. E. (2015). Neonicotinoid pesticide exposure impairs crop Potts, S. G., Biesmeijer, J. C., Kremen, C., Neumann, P., Schweiger, O., pollination services provided by bumblebees. Nature, 528, 548–552. & Kunin, W. E. (2010). Global pollinator declines: Trends, impacts https://doi.org/10.1038/nature16167 and drivers. Trends in Ecology & Evolution, 25, 345–353. https://doi. Tscharntke, T., Karp, D. S., Chaplin-Kramer, R., Batáry, P., DeClerck, org/10.1016/j.tree.2010.01.007 F., Gratton, C., … Zhang, W. (2016). When natural habitat Power, A. G. (2010). Ecosystem services and agriculture: Tradeoffs and fails to enhance biological pest control – Five hypotheses. synergies. Philosophical Transactions of the Royal Society B-­Biological Biological Conservation, 204, 449–458. https://doi.org/10.1016/j. Sciences, 365, 2959–2971. https://doi.org/10.1098/rstb.2010.0143 biocon.2016.10.001 Primdahl, J., Peco, B., Schramek, J., Andersen, E., & Onate, J. J. (2003). Williams, N. M., & Kremen, C. (2007). Resource distributions among Environmental effects of agri-­environmental schemes in Western habitats determine solitary bee offspring production in a mo- Europe. Journal of Environmental Management, 67, 129–138. https:// saic landscape. Ecological Applications, 17, 910–921. https://doi. doi.org/10.1016/S0301-4797(02)00192-5 org/10.1890/06-0269 Ramirez, F., & Davenport, T. L. (2013). Apple pollination: A review. Wratten, S. D., Gillespie, M., Decourtye, A., Mader, E., & Desneux, N. Scientia Horticulturae, 162, 188–203. https://doi.org/10.1016/j. (2012). Pollinator habitat enhancement: Benefits to other ecosys- scienta.2013.08.007 tem services. Agriculture Ecosystems & Environment, 159, 112–122. Reganold, J. P., Glover, J. D., Andrews, P. K., & Hinman, H. R. (2001). https://doi.org/10.1016/j.agee.2012.06.020 Sustainability of three apple production systems. Nature, 410, 926– 930. https://doi.org/10.1038/35073574 Rundlöf, M., Nilsson, H., & Smith, H. G. (2008). Interacting effects How to cite this article: Samnegård U, Alins G, Boreux V, et al. of farming practice and landscape context on bumblebees. Management trade-­offs on ecosystem services in apple Biological Conservation, 141, 417–426. https://doi.org/10.1016/j. orchards across Europe: Direct and indirect effects of organic biocon.2007.10.011 production. J Appl Ecol. 2019;56:802–811. Rusch, A., Birkhofer, K., Bommarco, R., Smith, H. G., & Ekbom, B. (2014). Management intensity at field and landscape levels affects the struc- https://doi.org/10.1111/1365-2664.13292 ture of generalist predator communities. Oecologia, 175, 971–983. https://doi.org/10.1007/s00442-014-2949-z