Ecological and evolutionary effects of crop diversity decrease yield variability

Improving the year-to-year yield stability of agroecosystems is fundamental under ongoing global change. While weather variability main cause variations in crop production (Osborne & Wheeler, 2013; Ray et al., 2015), climate models anticipate an increment frequency and intensity hazards (Pörtner 2022). Furthermore, commercial crops are bred to optimize current climatic conditions. Thus, these homogeneous varieties fail cope with uncertainty (Kahiluoto 2019). Consequently, efforts made develop agricultural practices that strengthen capacity for adaptation change decrease over time. Increasing diversity through intercropping (i.e. simultaneous cultivation more than one species on same field) has been suggested as effective practice sustainably promote (Gaudin 2015; Li 2021; Raseduzzaman Jensen, 2017; Renard Tilman, 2019; Ryan, 2021). Theoretical empirical work proposes temporal biomass plant communities increases higher richness (Tilman 2006; Wagg 2017). This suggests acts insurance against environmental changes sustains stable primary productivity time (Isbell Yachi Loreau, 1999). The stabilizing effect can result from several ecological mechanisms, being asynchrony fluctuations between (Valencia, de Bello, Galland, 2020). Increased means abundance some compensated by others. A pattern associated increased among could species-specific responses (Lepš 2018). Therefore, systems, diversification potentially lead probability including different functional strategies or adapted conditions—that is, asynchrony, maintaining compensating losses when others fail. Temporal demonstrated at national level (Egli Also, field level, combinations cereals legumes led lower sole (Weih So, stabilize even if not grown together field, this does require interact. Beyond differences preferences species, additional driver plant–plant interactions (competition facilitation) changing conditions (Bertness Callaway, 1994; Callaway 2002; He Michalet 2015). In case, do fluctuate independently each other a changes, but compensatory dynamics arise asymmetric competition For example, benign trigger hierarchical competitive thereby increasing 1998). Conversely, positive such facilitation due existence correlations benefactors beneficiaries. However, also play significant role buffering extreme (Mulder 2001). Benefactor moderate local environment climate, soil, etc.) be beneficial many allowing their survival performance across 2001; Wilby Shachak, 2004). shifts negative resource availability affect accentuating reducing species‘asynchrony (Butterfield, 2009). These mechanisms interact so exist fields. Evolutionary processes have recognized factor playing major ecosystem functioning, particularly fomenting effects. It well known using decreases abiotic biotic stresses (Villa 2005; Zeven, One goals breeding programs obtaining genotypes conditions, resilient (Newton 2011). mixed cropping it relevant find ‘cooperative’ reduce (Wuest become critical asymmetry species. An evolutionary mixtures address issue. Recent findings experimental grassland showed modification traits after generations mixture history) (Zuppinger-Dingley 2014). Inline, van Moorsel al. (2021) found long-term biodiversity experiment joint co-occurrence history decreased comparison naïve communities. selection yields Understanding factors driving may help design sustainable systems able maintain fluctuating environment. study, we evaluated monocultures (yield biomass), annual mixtures. We hypothesized that: (1) biomass) observed asynchronous compared monocultures; (2) reduced both given expected stressful (3) unfertilized fertilized marked Spain Switzerland drier growth season. conducted common garden Torrejón el Rubio (Cáceres, Spain) Zurich (Switzerland) three years (2018–2020). gardens were located Irchel campus (University Zurich: 47° 23′ 46″N, 8° 33′ 03″ E, 508 m a.s.l.) research station Aprisco (39° 48′ 48″N, 6° 00′ 01″ W, 350 a.s.l.). two sites represent soil semi-arid Mediterranean while temperate climate. Mean temperature total precipitation during growing seasons varied 14.5 16.7°C 63 326 mm 16.1 18.2°C 347 511 Switzerland, respectively. first year cooler wetter Spain, was warmer second third (Figure S1). Climatic data period (2018–2020) downloaded function get_daily_climate() easyclimate R package (Cruz-Alonso 2023; Moreno Hasenauer, 2016; Rammer plots squares 0.25 m2 raised beds around 35 cm depth. placed drained open bottom allow unlimited root growth. organized 15 10 × 1 16 7 Switzerland. Plots separated below-ground metal frames (35 deep). filled standard, enriched, coming region. Spanish (78% sand, 20% silt, 2% clay; pH 6.3; C N 0.5% 0.05%, respectively) sandier less fertile (45% 45% 10% 7.25; 3.39% 0.19%, respectively). irrigated ensure drought periods. automated irrigation system configured dry threshold moisture 17% capacity, target 25%. set 50% 90%. Whenever thresholds reached (measured PlantCare sensors (PlantCare Ltd., Switzerland)), initiated until reaching value. selected seven species: Avena sativa (oat), Triticum aestivum (wheat), Lens culinaris (lentil), Lupinus angustifolius (blue lupin), Camelina (camelina), Linum usitatissimum (linseed) Coriandrum sativum (coriander). because they exhibit similar phenology, requirements size, easily cultivated Europe present phylogenetic characteristics. classified four phylo-functional groups. Specifically, monocots (A. T. (Poaceae)); within dicots, superasterid (C. (Apiaceae)) superrosids, (L. L. (Fabaceae)) non-legumes (Brassicaceae) (Linaceae)). used locally country (the list cultivars suppliers Table whenever possible, traditional ancient open-pollinated order maximize genetic needed occur. applied treatment levels: monocultures, two-species four-species 1). include all distinct groups, study included 18 8 mixtures, replicates compositions. Several resulting possible currently intercropped without apparent economic interest. 2019 2020, additionally ‘selection history’ monoculture versus selection. history, sown communities, consisted assembled offspring plants had previous identical composition To grow seeds pool available (2018), original provided seed suppliers, (2019) (2020), harvested our own corresponding history. replicated set-up (factor ‘country’) fertility levels (non-fertilized control plots; ‘fertilization’) half nitrogen (N), phosphorus (P) potassium (K) concentration 120 kg/ha N, 205 P, K (ORGAMAX 7-12-7, Productos Agricolas MACASA S.L.): 41.5% fertilizer before sowing, wheat tillering stage, rest flowering. served controls. 2018, randomly allocated individual non-fertilized treatment. following years, kept initial fertilization allocation. Monoculture randomized year, combination replicating countries contrasting along treatment, allowed testing how shape diversity-stability relationship. planted 2 5 February, April, year. plot, rows between-row distance 12 cm. Each assigned planting row plot. sowed hand standard sowing density depth (see Harvest maturity. Seeds sun-dried days weighed. determine vegetative biomass, clipped right above surface. Vegetative stems, leaves chaff, oven-dried 80°C 48 h weighing. (seed yield) above-ground measures productivity. tested community- species-level. At community-level, mass community. species-level, multiplied number community account fact densities depending compare calculated generated monocultures. average (same fertilization, country). then summed productivities divided adjusted coefficient variation (variabilityaCV). VariabilityaCV which removes dependence variance mean. separately type community, country, composition, Döring Reckling (2018) acv metan (Olivoto Lúcio, noteworthy comparing mixtures), rule out potential “portfolio effect” mechanism responsible relationship (Doak 1998; Mccann, 2000). yield-stabilizing effects would mathematical artefact derived statistical averaging, predict sum independent progressively summed. assessed variabilityaCV (crop diversity, level). evaluate variability, (selection level) there (2019 2020) 2018. 2018 calculate mean response (at species-level) linear mixed-effects models. (one, species), (2018 2020), (yes, no) fixed (Spain, Switzerland) thresholds, starting material. meet model assumptions normality homoscedasticity errors, root-transformed coded heteroscedastic structure (‘weights’ argument ‘lme’ function). (expected observed), effects, instead diversity. random effect, mixture, effect. performed separate test (monoculture vs mixture) its variabilityaCV, analysis excluded suppliers. Besides (monoculture, mixture), (2019, Switzerland), Species Differences treatments analysed detail Tukey's post-hoc comparisons. removed data, where least plot no did specified zero values maintained long 101 (2,7%) 3697 species-level samples, 59 (3,4%) 1671 plot-level samples. significantly 45.0% 51.5% Figure S2; S2). marginal 2), 15.7% 9.6% 4.4% 4.7% 19.1% differ (Table that, S3, 3). affected S4). Post-hoc comparisons revealed (p = 0.002), 0.363; 4A). soils 4B). Regarding any difference species- nor community-level) despite (Figures S3 As expected, 5A). dependent indicated interaction S5). tests exhibited 5B). Biomass S5, S6). Community-level (monocultures vs. mixtures; Pairwise productive words, 22.2% 0.013, 6) 15.5% 0.021, S6) treatments, S7). pairwise composed (and only Our results provide evidence (i) community-level (ii) increase decreasing (iii) strongly (although certain conditions) provides yield. facilitative tends 2009; Mulder greater stronger influence facilitation, shown natural grasslands 2018), indicate pronounced Similarly, line showing field-level idea intercrop exhibiting supported results, reflect covariation stability. Taken together, synchrony suggest interspecific foster intensified 2018; 1996). mineral fertilizers intensify production. mixture. necessarily fact, likely overrun underlying rationale types originating display rapid transgenerational surrounding Indeed, statement Stefan (2022) who leaf matter content accordance studies variety high diversity) tend (Kiær Reiss Drinkwater, Wuest demonstrate approaches improve intraspecific niche complementarity usually (Chen Isbell Because stabilizes based 2006), larger (Loreau Mazancourt, Prieto result, yields, act selective force yields. Stomph supporting general (Hector 1999; Tilman countries, indicating benefit context dependent. contrast experiments maize (Li behind biodiversity–productivity relationship, sampling interannual (Barot Engbersen Annual well-known therefore, determinant (Moore Lobell, large benefits coincide temperatures registered. seem coupled S1), highlighting context-dependence systems. Despite years. caused Such hierarchies variability. favouring species' favour ability unfavourable consistent (Tilman, 1996; Ruijven Berendse, 2007). Compensatory coexisting explain patterns community-levels. There caveats should noted. First, detailed results. often case Valencia, Lepš, Here, furthermore, combination. multifactorial offers advantages controlling generally might compensate use 3 estimate important note last assessing since first-year newly purchased variability). must mind considering Finally, dependent, becoming most harsh water-limited environments Spain). explained keep probable wild genetically heterogeneous histories exert pressures, summary, vary context. diverse harsher implications enhancing face rising temperatures. More investigate affecting show driven dynamics, particular compromised. farmer's perspective, specific key, favoured intercrops, intercrops cash together. short-term suggesting promotes Current largely focused inclusion multiplication schemes consequently differentiation. Christian Schöb conceived designed Laura Nadine Engbersen. All authors collected data. Jesús López-Angulo analytical framework, writing manuscript. comments manuscript gave final approval publication. thank Elisa Pizarro, Carlos Barriga, Sandra González Sánchez, Ingala, Anna Bugmann, Anja Schmutz Lukas Meili assistance. funded Swiss National Science Foundation (PP00P3_170645 PP00P3_198906). Open access funding Eidgenossische Technische Hochschule Zurich. declare conflict peer review article https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/1365-2745.14092. Dataset code repository Zenodo https://doi.org/10.5281/zenodo.7544811 (López-Angulo S1. List depth, ecotypes Spain. S2. Results ANOVA (type II) interactions. NumDF, degrees freedom term; DenDF, error F-value, ratio; probability. * (P < 0.05), ** 0.01), *** 0.001). S3. S4. S5. S6. S7. (a) (˚C), (b) monthly (˚C) (c) February June April August Error bars errors. (in g m−2) (2018, (Spain Switzerland). Points 95% confidence intervals, ANOVAs presented (a; (b; yes, fertilized; no, fertilized) (c; 2019, Shared letters (Tukey's test, 0.05). fertilized). mixture). history), mixtures) (yes no). alpha Please note: publisher functionality information supplied authors. Any queries (other missing content) directed author article.