Weed communities are more diverse, but not more abundant, in dense and complex bocage landscapes

Through landscape simplification, decreased crop diversity and increased agrochemical use, agricultural intensification led to a dramatic loss of weed (Fried et al., 2009; Meyer 2013), evolution herbicide resistance in weeds (Neve 2009). Since 2010, around 3 million tons herbicides are used each year worldwide eliminate species (FAO, 2021), most cases regardless their actual impact on yield ecosystem service provision. However, only small proportion causes significant (Adeux 2019; Marshall 2003). For example, 306 main described Germany (Hofmeister & Garve, 1998), 26 were defined as problematic. On the contrary, can be major asset for biodiversity conservation provision agroecosystems (Marshall There is growing evidence that maintaining enhances multifunctionality sustainability (Gaba 2020; Liebman 2021). In addition, key agroecological management. First, diversified communities provide wider range habitat trophic resources natural enemies, which promotes biological control themselves Schumacher 2020). Second, diversification prevent dominance competitive limit due weed-crop competition Dornelas Storkey Neve, 2018), likely by promoting greater niche differentiation resource partitioning between crops (Smith 2010). seminal review, MacLaren al. (2020) emphasized selects fast-growing tall with ruderal strategies. These authors suggested reduced availability management practices habitats should promote more diverse life Hedgerows may play role this regard, but effects remain understudied, especially at scale. many services, including sustainable food biomass production, soil water protection, carbon sequestration, (Montgomery other hand, farmers often concerned hedgerows harbour plants spreading into arable fields (Alignier relatively stable support high plant (Vanneste Thus, spillover from could increase cover fields. Under conventional farming, typically restricted field edges, limited dispersal distances regular disturbances tillage operations use (Figure 1a) (e.g. Marshall, 1989; Metcalfe 2019). Plant extend over larger areas under organic farming given absence chemical fertilizer treatments 2022), studies lacking. Further, associated an ‘transient’ (Metcalfe not typical ‘resident’ emerge permanent seedbank, rely adjacent source habitats. As such, transient adapted lower characterized or stress-tolerant strategy. influence through another overlooked process. affect microclimate wind speed, light, temperature, frost), parameters matter, nitrogen flows), cycle atmospheric moisture, flows) 1b) Forman Baudry, 1984; Sánchez intensity hedgerow varies distance hedgerows, features (e.g., dimensions, permeability), orientation spatial configuration landscape. This result degree environmental heterogeneity fields, providing broader ecological conditions species. Niche would then allow coexist reduce (Dornelas Most assessing semi-natural Alignier 2017; Dainese ‘bocage’ vary systems (conventional vs. organic, scales). one expect pronounced context, germination growth (Gabriel 2010; where already levels (Roschewitz 2005). Our objective was assess (i.e. density complexity networks) cores, investigate potential interactions scales. addition taxonomic approach, we functional approach whether bocage landscapes mitigate selection strategies, allowing co-existence strategies communities. The leaf-height-seed scheme proposed Westoby (1998) very useful describe aspects response disturbances. Specific leaf area (SLA) related fast versus slow corresponding axis Grime's CSR scheme, opposing (high SLA) stress tolerance strategy (low SLA). height seed mass reflect separate coping competition, opposition stature large seeds) (short hand. showed specific area, height, along gradients precipitations, nutrient availability, reviewed Garnier (2016). We evaluated contribution cores. ‘plant spillover’ hypothesis, well ‘environmental heterogeneity’ diversity, cover, while reducing weeds. study conducted southern part Zone Atelier Armorique, Long-Term Socio-Ecological Research (LTSER) site Brittany, northwestern France (47°59′35N, 1°45′12W). region dense networks crop-livestock systems. Agricultural lands dominated wheat maize temporary grasslands, dominant types Brunisols Luvisols 2). mostly old planted least before World War II), generally composed oak Quercus robur chestnut Castanea sativa trees bank pruned firewood every 9–12 years. When present, shrub layer hazel Corylus avellana, hawthorn Crataegus monogyna, blackthorn Prunus spinosa, spindle Euonymus europaeus, broom Cytisus scoparius gorse Ulex europaeus. A detailed description based sample 40 provided Boinot (2022). Photographs illustrating our Figure S1. Observation sites (n = 37) located two independent (r 0.15, p-value 0.20): total length (ranging 6.3 18 km within buffer radius 1000 m focal fields) 0% 37% m) Each contains pair less than 1 apart managed (CF OF). Among these 62 grown winter cereals, whereas 12 OF cereal-legume mixtures (we resorted when find any pure area). Management similar S2). All gave us permission fieldwork. did require ethical approval. field, sampled ten m2 core, 20 margin (sampling designs data set S3). pairs 2019 17 2020, June July harvest. Pairs different those evenly same gradients. visually estimated found quadrats, accuracy ±5%. Ten quadrats per necessary obtain fairly accurate estimate indicated accumulation curves S4). calculated mean (%) 10 field. Then, measured (i.e., cumulative all species), (Hill-Shannon index) Hill-Shannon considers both number community relative calculate rarity (Roswell no CF therefore assigned lowest possible value (zero). To CWM CWV Ellenberg indicator values temperature (EIV-T), moisture (EIV-F), light (EIV-L), nutrients (EIV-N). An indicates higher heterogeneity. shift suggests change conditions. collected Baseflor database (Julve, 1998). Species known always represented 90% Correlations traits values, fidelity scores 0.4 S5). Generalized plots variables Figures S6 S7. assessed local treatment, N-fertilization (Figures S12 S13). considered mineral fertilization, converting fertilization amounts hectare. Maize—winter far common rotation systems, have complex rotations, polycultures grasslands. average, surveyed been 23 ± 6 Finally, computed (%OF) circular radii 250, 500, 750, geographical Department Ille-et-Vilaine (provided CartoBio team French Agency Development Promotion Organic Agriculture). generalized additive models (GAMs) effect ‘bocage’, system scales, two-way cores 74 fields). coordinates included GAMs account autocorrelation, thanks smooth term modelling interaction latitude longitude 2018). ‘Year’ also covariate. Gamma indices, Seed mass, Gaussian using r package mgcv (Wood, 2017). added constant (10−6) 0 GAMs. SLA, EIV-L, removed extreme unduly influencing results. explanatory standardized (mean-centred divided standard deviations) compare importance (Schielzeth, collinearity checked analyses S14–S17) variance inflation factors (VIF). had VIF thus full (Zuur 2013). proceeded model averaging procedure information-theoretic (Burnham 2011), MuMIn (Barton, 95% confidence summed Akaike weights largest smallest until reaching 0.95) delineate top set. performed analysis scale separately. Across surveyed, recorded 125 belonging 28 families. Relative families decreasing order: Poaceae (20%), Fabaceae (19%), Juncaceae (16%), Asteraceae (11%), Caryophyllaceae (4%), Primulaceae Plantaginaceae Papaveraceae (3%), Convolvulaceae (3%) Rosaceae (3%). Top frequent observed Juncus bufonius (68%), Lysimachia arvensis (61%), Poa annua (59%), Veronica (58%), Polygonum aviculare (55%), Trifolium repens (51%), Tripleurospermum inodorum (45%), Viola (43%), Convolvulus (43%) Aphanes (42%). More information composition OF) S18. ‘Bocage’ whatever 3a S20). Total (60 30% 9 14% CF). communities, 3b S21). Higher %OF tended (12 5 4 CF) affected community-weighted (CWM) (CWV) S22–S27). 250 (and almost significantly 500 S24). SLA 3d S23), 3f S25), 3h S27), 750 m. except (from S23). S27). Regarding Fidelity, 4a S28). Fidelity there negative 4b S29). (EIV-L) multiple scales 5, S31, S33 S35). (EIV-N) S37), 5h). EIV-T EIV-L 5b,f). S33, S35 S37). contrast CWV, influenced results consistent across S30, S32, S34 S36). Nonetheless, EIV-N m, 5g (2017), who cover. Weed strongly depends rather context (Petit 2016). Indeed, variable affecting strong positive treatments, shown Gabriel Roschewitz (2017) richness, diversity. Hedgerow densities, studies. difference arise regional contexts climate, management, history) structures, compositions, configurations simple single-storied north-eastern Italy heterogeneous Brittany). choice indices explain contrasting results; considering Surprisingly, scale, enhanced despite stronger line previous studies, tends (Henckel 2015; Petit 2016), Yet, probably because favoured share attributes. mass. suggest drives landscapes. contrary what expected hypothesis processes, Undoubtedly, simplified rotations pressures phenological traits) Rotchés-Ribalta homogenization caused destruction has impact, largely restricts Results non-crop further hypothesis. habitats, indicating towards (at nearest hedgerows) anecdotal non-existent. showing detrimental edge, Raatz Increased presence sown grasslands vetches clovers). Additional come such herbaceous margins undisturbed act physical barriers preventing sink unfavourable pedoclimatic shade) (Wilkerson, 2014). lesser extent nutrients, landscapes, suggesting preponderant assembly co-existence. instance, García De León (2014) demonstrated changes spring balance species, depending preferences rates. Competition determinant (Perronne Besides, Silvertown (1999) revealed fine-scale variation allows meadow hold Beyond abiotic factors, enhance biotic turn favouring requires investigation. pollinators (Morandin Kremen, dispersers (Heath 2017), predators (Badenhausser 2020) (Fontaine 2006; Hulme, 1998; Ozinga Contrary expectations, synergetic Alternatively, assumed reach carrying capacity even intensive 2005), case either. meta-analysis, Seiferling particularly beneficial highly disturbed homogenized ecosystems. increasing biodiversity, adverse effects, Arable ecosystems, organically, annual rotations. increases opportunities way (regardless system). coexistence, indeed adequate take advantage opportunities. found, well-preserved highlighting cannot if too strong. Farmers sometimes spray directly base hedge flora growing. weeds, Therefore, microhabitat farmland (Gardarin causing losses (Storkey additional argument reintroduction shrubs deliver services conservation, sequestration. exciting avenue research quality improve understanding production. survived being degraded inappropriate neglect 2015), like European regions (Deckers 2004; Litza Diekmann, Tilzey, lose structure microclimate, undermines functions livestock sequestration). preserve functionality include coppicing hedge-laying, reduction flailing frequency, preservation canopies intervals, protection browsing domestic animals drifts (Tilzey, Following guidelines, network favour conservation. Audrey Alignier, Stéphanie Aviron Cendrine Mony conceived ideas designed methodology; Claire Ricono, Eloïse Couthouis data; Sébastien analysed writing manuscript; Mony, Guillaume Fried, Aude Ernoult contributed drafts. final approval publication. La Fondation de (DIVERSIT'HAIES project) financially supports postdoctoral SB. project received financial Office Français pour la Biodiversité (AGRIM Armorique (DIVAG project). thank Gérard Savary farmers' interviews, Noémie Poulain Sarah Leclerc surveys, KERMAP mapping, Agriculture farming. associate editor, Pieter Frenne, anonymous reviewers constructive comments earlier version manuscript. None. Data available via Dryad Digital Repository https://doi.org/10.5061/dryad.k0p2ngfbv (Boinot 2022). Appendix S1 Please note: publisher responsible content supporting supplied authors. Any queries (other missing content) directed author article.