A small effect of conservation agriculture on soil biodiversity that differs between biological kingdoms and geographic locations

•Agrochemicals have a small effect on soil bacterial, fungal, and animal biodiversity•The differs in space (and less so time) affects taxa abundances greatest•There is no of agrochemicals richness•The size the between but greatest for eukaryotic Larger easily visible animals plants are negatively affected by used intensive food production, we do not understand general spatial temporal effects multitudes bacteria, fungi, invertebrate that underpin ecosystem productivity. We sequenced 16S, ITS2, COI DNA barcode regions from 648 New Zealand vineyard samples managed under either conventional or low-agrochemical-input conservation approaches across two three seasons 1 year discovered at least 170,000 phylotypes (taxa) with >97% genetic identity. Management approach correlated significant 2%–10% difference differed over seasons. Although data show large biodiversity average, important finding magnitude impact types locations, some most also influence quality agricultural produce. It well documented pesticides agriculture decrease to naked eye (Altieri, 1999Altieri M. The ecological role agroecosystems.Agric. Ecosyst. Environ. 1999; 74: 19-31Crossref Scopus (1454) Google Scholar; Bengtsson et al., 2005Bengtsson J. AhnströM Weibull A. organic abundance: meta-analysis: agriculture, abundance.J. Appl. Ecol. 2005; 42: 261-269Crossref (1004) Billeter 2008Billeter R. Liira Bailey D. Bugter Arens P. Augenstein I. Aviron S. Baudry Bukacek Burel F. al.Indicators landscapes: pan-European study.J. 2008; 45: 141-150Crossref (462) Clough 2005Clough Y. Kruess Kleijn Tscharntke T. Spider diversity cereal fields: comparing factors local, landscape regional scales.J. Biogeogr. 32: 2007-2014Crossref (175) Gabriel 2010Gabriel Sait Hodgson Schmutz U. Kunin W. Benton Scale matters: farming different scales.Ecol. Lett. 2010; 13: 858-869Crossref PubMed (275) Gonthier 2014Gonthier Ennis K. Farinas Hsieh H. Iverson Batáry Rudolphi Cardinale B. Perfecto Biodiversity requires multi-scale approach.Proc. Soc. B Biol. Sci. 2014; 281: 20141358Crossref (173) McLaughlin Mineau, 1995McLaughlin Mineau practices biodiversity.Agric. 1995; 55: 201-212Crossref (413) Puig-Montserrat 2017Puig-Montserrat X. Stefanescu C. Torre Palet Fàbregas E. Dantart Arrizabalaga Flaquer Effects crop management 2017; 243: 19-26Crossref (36) Rundlöf Smith, 2006Rundlöf Smith butterfly depends context.J. 2006; 43: 1121-1127Crossref (191) Scholar), these represent only tiny fraction global biodiversity. Soils harbor one-quarter world's biodiversity, approximately 40% globe's land area dedicated agriculture. However, vast array animals, which productivity ecosystems, particularly via soils, poorly characterized (Guerra 2020Guerra Heintz-Buschart Sikorski Chatzinotas Guerrero-Ramírez N. Cesarz Beaumelle L. Rillig Maestre Delgado-Baquerizo al.Blind spots function research.Nat. Commun. 2020; 11: 3870Crossref (43) Harkes 2019Harkes Suleiman van den Elsen de Haan Holterman Kuramae Helder Conventional as divergent drivers resident active fractions major web constituents.Sci. Rep. 2019; 9: 13521Crossref (20) Hartmann 2015Hartmann Frey Mayer Mader Widmer Distinct microbial long-term farming.ISME 2015; 1177-1194Crossref (601) Morrison-Whittle 2017Morrison-Whittle Lee Goddard M.R. Fungal communities differentially biodynamic ecosystems.Agric. 246: 306-313Crossref (46) Scholar). Greater positively correlates increased stability, function, resilience, nutrient recycling, detoxification, pest control, decreased requirement fertilizer pesticide input agroecosystems (Awasthi 2014Awasthi Singh Soni Kalra acts insurance bacterial abiotic perturbations.ISME 8: 2445-2452Crossref (78) Blanchet 2016Blanchet G. Gavazov Bragazza Sinaj Responses properties yields inorganic amendments Swiss system.Agric. 2016; 230: 116-126Crossref (73) 2012Cardinale Duffy Gonzalez Hooper Perrings Venail Narwani Mace Tilman Wardle al.Biodiversity loss its humanity.Nature. 2012; 486: 59-67Crossref (3365) Harrison 2014Harrison Berry Simpson Haslett Blicharska Bucur Dunford Egoh Garcia-Llorenre Geamānā al.Linkages attributes services: systematic review.Ecosyst. Serv. 191-203Crossref (328) McCann, 2000McCann diversity-stability debate.Nature. 2000; 405: 228-233Crossref (1872) McGrady-Steed 1997McGrady-Steed Harris P.M. Morin regulates predictability.Nature. 1997; 390: 162-165Crossref (539) Médiène 2011Médiène Valantin-Morison Sarthou Tourdonnet Gosme Bertrand Roger-Estrde Aubertot Rusch Motisi al.Agroecosystem biotic interactions: review.Agron. Sustain. Dev. 2011; 31: 491-514Crossref (112) Naeem 1995Naeem Thompson Lawler Lawton Woodfin Empirical evidence declining phylotype may alter performance terrestrial ecosystems.Philos. Trans. 347: 249-262Crossref (293) Paoletti, 2001Paoletti M.G. bioindicators environmental health.in: Shiyomi Koizumi Structure Function Agroecosystem Design Management. CRC Press, 2001: 11-38Google 2014Tilman Isbell Cowles J.M. functioning.Annu. Rev. Evol. Syst. 471-493Crossref (765) There correspondingly increasing grower consumer interest “conservation” approaches, consider longer-term wider impacts agriculture: agrochemical inputs an attempt increase (Bommarco 2013Bommarco Potts S.G. Ecological intensification: harnessing services security.Trends 2013; 28: 230-238Abstract Full Text PDF (879) Ponti 2012de Rijk Ittersum M.K. yield gap agriculture.Agric. 108: 1-9Crossref (520) Döring 2019Döring Collins Frisch Kauer Organic viticulture affect vine wine: quantitative review.Am. Enol. Vitic. 70: 221-242Crossref (14) Matson 1997Matson P.A. Parton W.J. Power A.G. Swift M.J. Agricultural intensification properties.Science. 277: 504-509Crossref (1982) 2001Tilman Fargione Wolff D’Antonio Dobson Howarth Schindler Schlesinger Simberloff Swackhamer Forecasting agriculturally driven change.Science. 2001; 292: 281-284Crossref (2507) studies date evaluated larger (Bengtsson massively more abundant complex micro- mesofauna time described. analysis fungal amplification sequencing millions 16S ITS barcodes directly extracted now commonplace circumvents problems associated fact >95% species missed culture-based methods (Taylor 2014Taylor M.W. Tsai Anfang Ross H.A. Pyrosequencing reveals differences fruit-associated communities.Environ. Microbiol. 16: 2848-2858Crossref (105) few estimates evaluate had limited sampling locations points suggested approach, sizes relatively small. Scholar showed one site (∼10%) (10%) multiple vineyards region point (Morrison-Whittle reported (∼4%) managements farms using eukaryote 18S barcodes, Bonanomi 2016Bonanomi De Filippis Cesarano La Storia Ercolini Scala induces changes microbiota agro-ecosystem functions.Soil Biochem. 103: 327Crossref (70) similar polytunnel systems same (but did report sizes). aware standard barcode. Evidence broader components lacking, research quantifies commercially relevant both will provide inform policy this regard (Godfray 2010Godfray H.C.J. Beddington Crute I.R. Haddad Muir J.F. Pretty Robinson Thomas S.M. Toulmin Food security: challenge feeding 9 billion people.Science. 327: 812-818Crossref (5938) Zimmerer Haan, 2017Zimmerer K.S. Agrobiodiversity sustainable future.Nat. Plants. 3: 1-3Crossref (31) Here study commercial sites operate approaches. deliberately chose range sites, rather than experimental plots, quantify authentic scenarios. gathered 24 located 350 km apart Marlborough Hawke's Bay (HB) spring, summer, autumn. tested whether there were multi-kingdom regimes analyzing then went estimate nature any put into context. Analysis spray diaries (detailed Table S1) revealed 25% application events 3-fold significantly greater products per hectare period sampled (Mann-Whitney U test, Z = 6.41, n 574, p 1.4 × 10−10): substantiates classifications sites. One million seven hundred thousand sequence reads obtained soils after forward-reverse pairing filtering (561,409 16S; 443,082 ITS2; 724,661 COI). All sequences clustered 97% identity, level separates prokaryote genera (Alberdi 2017Alberdi Aizpurua O. Gilbert M.T.P. Bohmann Scrutinizing key steps reliable metabarcoding samples.Methods 134-147Crossref (219) Guerra Hebert 2003Hebert P.D.N. Ratnasingham deWaard J.R. Barcoding life: cytochrome c oxidase subunit divergences among closely related phylotype.Proc. 2003; 270: S96-S99Crossref (3219) Konstantinidis Tiedje, 2005Konstantinidis K.T. Tiedje Genomic insights advance definition prokaryotes.Proc. Natl. Acad. S 102: 2567-2572Crossref (1190) Twelve comprising 133,575 matched Homo genus removed all further analyses leaving 172,370 1.5 106 (116,788 2,557 53,025 phylotypes), first approximation total ecosystems (Table S2). desirable taxonomically classify matching their those reference databases. Eighty-five percent probabilistically assigned kingdom, 96% ITS2 kingdom. Actinobacteria was phylum (29%), followed Proteobacteria (26%), ascomycetes dominated fungi (79%): complete breakdown taxonomic levels S2. identification challenging databases far complete. Comparisons (see methods) 99.99% eukaryotes, half deposits classified kingdom confidence. Approximately 3% (metazoans), ∼0.3% oomycetes fungus-like organisms. recovered yet databases, does prevent analyses, just assignment. Eukaryotes contain regions, due targeting PCR primers. thus generally. analyzed each separately, completeness combined barcodes. be heavily influenced number phylotypes, it possible overlap barcodes; therefore primarily focused individually. Rank abundance curves typical patterns common many rarer (Figure 1): bacteria richest evenly distributed, eukaryotes fungi. differential abundances, presences, counts comprises core being able insensitive normality, variance, sample sizes. conservatively sub-sampled (rarefied) 2,000 1,000 ensure equal within barcode; fewer analyses. Following rarefaction 275,000 79,047 S3). null hypothesis significance when rejected. results including test statistics, Tables S4, S5, S6, S7, (the proportion variance explained factor) text < 0.05 brevity. consistent combination) 2; S4). This weaker (p 0.022 0.013, R2 0.019 0.051), stronger 0.005, 0.018). interactions individual 0.09 0.43), HB 0.011 0.020, 0.045 0.105) 0.760 0.185). phylotypes: summer 0.002, 0.202 0.028, 0.055 respectively), season Phylotype irrespective 0.001; S4; Figure 2). 0.001 mean 0.08), 0.296). Where significant, (5.6% versus 4.8% 8.1% season). presences 0.018, 0.001, 3), 2% (R2 0.018; S5). 0.121), and/or > 0.3). marked interaction 0.068), whereas 0.04) 0.646), weak autumn 0.048, 0.054). 0.187): 0.038 0.017; 0.034 0.033). 0.002 0.02; 3; 0.004. 0.057), 0.11). 2.9% almost identical (2.8%). (counts) present individually 0.1–0.86; 4; S6). weakly management, region, 0.021 0.010 translate separately 0.09–1). Shannon indices, account distribution counts, 0.11–0.92; S1). above-mentioned meaningful, they specific previous did. Bacterial greatly 0.003, E2 0.32), 0.76), where average 297 budburst harvest 0.008; 4). 0.027 0.0001; 0.116 0.23, respectively) ∼30% normalized entirety cumulative sum scaling, sensitive depths, mentioned earlier S7). depths rarefaction: medium ranges 0.03 0.001) apparent region. Given indicator employed probability proportions ways. forty-six (Padj values 0.05): 112 overrepresented 34 S8). Thermoleophilia Rubrobacteria classes conventionally 0.021), Bacilli (class) lower Firmicutes (phylum; corrected 0.033–0.05) vineyards. Seventeen approaches: eleven Bionectria, Malassezia, Saccharomyces, Mortierella (corrected 0.009–0.037, S8) six Ascomycete (phylum) Trichocladium unidentified indicative Thirty-four statistically fourteen As issues reliably identifying describe S9). attempted match nucleotide BLAST searches whole GenBank S10 see methods). 14 Mus musculus (mice) 99.4% another nine Mammalia 90%–96% identities, meaning highly likely mammals, probably Muridae (rodents). Two other ∼97% matches finer robust classification possible. last conservation-indicative has been described best ∼94% Hemiptera (insect, tree bugs) Ascomycetes (fungi). poor (83%–95%) various species, seems added detected raw non-rarefied Botryosphaeria spp., Eutypa lata, Phaeomoniella pathogens implicated canker esca-like trunk diseases, disease agents 0.072). No Erysiphe, Plasmopara, Botrytis genera, cause grape powdery downy mildew bunch rot, recovered, recall samples. Overall detectable soil-derived way 1), generally (<10%). contingent location lesser extent). strongest suggestion summer. HB, signal summer.Table 1The numbers, types, (HB)NumberTypeAbundance16S0.100HB - 0.005R2 0.04HB 0.011R2 0.045ITS20.3010.121HB 0.020R2 0.105COI0.8670.001R2 0.0180.005R2 0.019p shown derive Kruskal-Wallis (number PERMANOVA (types abundance) HB-specific indicated indicate component Marlborough. (R2)