Ecology

Sharon I, Kertesz M, Hug LA, Pushkarev D, Blauwkamp TA, Castelle CJ, et al. Accurate, multi-kb reads resolve complex populations and detect rare microorganisms. Genome Res. 2015;25:534–43.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Bankevich A, Pevzner PA. Joint analysis of long and short reads enables accurate estimates of microbiome complexity. Cell Syst. 2018;7:192–200.e3.CAS 
PubMed 
Article 

Google Scholar 
Luo C, Tsementzi D, Kyrpides NC, Konstantinidis KT. Individual genome assembly from complex community short-read metagenomic datasets. ISME J. 2012;6:898–901.CAS 
PubMed 
Article 

Google Scholar 
Lapidus AL, Korobeynikov AI. Metagenomic data assembly—the way of decoding unknown microorganisms. Front Microbiol. 2021;12:613791.PubMed 
PubMed Central 
Article 

Google Scholar 
Nielsen HB, Almeida M, Juncker AS, Rasmussen S, Li J, Sunagawa S, et al. Identification and assembly of genomes and genetic elements in complex metagenomic samples without using reference genomes. Nat Biotech. 2014;32:822–8.CAS 
Article 

Google Scholar 
Biller SJ, Berube PM, Lindell D, Chisholm SW. Prochlorococcus: the structure and function of collective diversity. Nat Rev Microbiol. 2015;13:13–27.CAS 
PubMed 
Article 

Google Scholar 
Crespo BG, Wallhead PJ, Logares R, Pedrós-Alió C. Probing the rare biosphere of the North-West Mediterranean Sea: an experiment with high sequencing effort. PLOS ONE. 2016;11:e0159195.PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Sogin ML, Morrison HG, Huber JA, Welch DM, Huse SM, Neal PR, et al. Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc Natl Acad Sci USA. 2006;103:12115–20.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Pedrós-Alió C. Dipping into the rare biosphere. Science. 2007;315:192–3.PubMed 
Article 

Google Scholar 
Sauret C, Séverin T, Vétion G, Guigue C, Goutx M, Pujo-Pay M, et al. ‘Rare biosphere’ bacteria as key phenanthrene degraders in coastal seawaters. Environmental Pollution. 2014;194:246–53.CAS 
PubMed 
Article 

Google Scholar 
Kalenitchenko D, Le Bris N, Peru E, Galand PE. Ultra-rare marine microbes contribute to key sulfur related ecosystem functions. Mol Ecol. 2018;27:1494–504.PubMed 
Article 

Google Scholar 
Capo E, Debroas D, Arnaud F, Guillemot T, Bichet V, Millet L, et al. Long-term dynamics in microbial eukaryotes communities: a palaeolimnological view based on sedimentary DNA. Mol Ecol. 2016;25:5925–43.CAS 
PubMed 
Article 

Google Scholar 
Lynch MDJ, Neufeld JD. Ecology and exploration of the rare biosphere. Nat Rev Micro. 2015;13:217–29.CAS 
Article 

Google Scholar 
Debroas D, Hugoni M, Domaizon I. Evidence for an active rare biosphere within freshwater protists community. Mol Ecol. 2015;24:1236–47.CAS 
PubMed 
Article 

Google Scholar 
Banerjee S, Schlaeppi K, Heijden MGA. Keystone taxa as drivers of microbiome structure and functioning. Nat Rev Microbiol. 2018;16:567–76.CAS 
PubMed 
Article 

Google Scholar 
Herren CM, McMahon KD. Keystone taxa predict compositional change in microbial communities. Environ Microbiol. 2018;20:2207–17.PubMed 
Article 

Google Scholar 
Hugoni M, Taib N, Debroas D, Domaizon I, Dufournel IJ, Bronner G, et al. Structure of the rare archaeal biosphere and seasonal dynamics of active ecotypes in surface coastal waters. PNAS. 2013;110:6004–9.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Debroas D, Domaizon I, Humbert J-F, Jardillier L, Lepère C, Oudart A, et al. Overview of freshwater microbial eukaryotes diversity: a first analysis of publicly available metabarcoding data. FEMS Microbiol Ecol. 2017;93:1.Article 
CAS 

Google Scholar 
Elshahed MS, Youssef NH, Spain AM, Sheik C, Najar FZ, Sukharnikov LO, et al. Novelty and uniqueness patterns of rare members of the soil biosphere. Appl Environ Microbiol. 2008;74:5422–8.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Pascoal F, Magalhães C, Costa R. The Link Between the Ecology of the Prokaryotic Rare Biosphere and Its Biotechnological Potential. Front Microbiol. 2020;11:231.PubMed 
PubMed Central 
Article 

Google Scholar 
Rinke C, Schwientek P, Sczyrba A, Ivanova NN, Anderson IJ, Cheng J-F, et al. Insights into the phylogeny and coding potential of microbial dark matter. Nature. 2013;499:431–7.CAS 
PubMed 
Article 

Google Scholar 
Delmont TO, Eren AM, Maccario L, Prestat E, Esen ÖC, Pelletier E, et al. Reconstructing rare soil microbial genomes using in situ enrichments and metagenomics. Front Microbiol. 2015;6:358.PubMed 
PubMed Central 

Google Scholar 
Sachdeva R, Campbell BJ, Heidelberg JF Rare microbes from diverse Earth biomes dominate community activity. bioRxiv 2019; 636373. https://doi.org/10.1101/636373.Galand PE, Pereira O, Hochart C, Auguet JC, Debroas D. A strong link between marine microbial community composition and function challenges the idea of functional redundancy. ISME J. 2018;12:2470–8.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30:2114–20.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Peng Y, Leung HCM, Yiu SM, Chin FYL. IDBA-UD: a de novo assembler for single-cell and metagenomic sequencing data with highly uneven depth. Bioinformatics. 2012;28:1420–8.CAS 
PubMed 
Article 

Google Scholar 
Li H, Durbin R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics. 2010;26:589–95.PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Ulyantsev VI, Kazakov SV, Dubinkina VB, Tyakht AV, Alexeev DG. MetaFast: fast reference-free graph-based comparison of shotgun metagenomic data. Bioinformatics. 2016;32:2760–7.CAS 
PubMed 
Article 

Google Scholar 
Dixon P. VEGAN, a package of R functions for community ecology. J Vegetation Sci. 2003;14:927–30.Article 

Google Scholar 
Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41:D590–D596.CAS 
PubMed 
Article 

Google Scholar 
Truong DT, Franzosa EA, Tickle TL, Scholz M, Weingart G, Pasolli E, et al. MetaPhlAn2 for enhanced metagenomic taxonomic profiling. Nature methods. 2015;12:902–3.CAS 
PubMed 
Article 

Google Scholar 
Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, et al. Metagenomic biomarker discovery and explanation. Genome Biology. 2011;12:R60.PubMed 
PubMed Central 
Article 

Google Scholar 
The UniProt Consortium. UniProt: the universal protein knowledgebase. Nucleic Acids Res. 2017;45:D158–D169.Article 
CAS 

Google Scholar 
Kanehisa M, Sato Y, Kawashima M, Furumichi M, Tanabe M. KEGG as a reference resource for gene and protein annotation. Nucleic Acids Res. 2016;44:D457–D462.CAS 
PubMed 
Article 

Google Scholar 
Buchfink B, Xie C, Huson DH. Fast and sensitive protein alignment using DIAMOND. Nat Meth. 2015;12:59–60.CAS 
Article 

Google Scholar 
Fernandes AD, Reid JN, Macklaim JM, McMurrough TA, Edgell DR, Gloor GB. Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective growth experiments by compositional data analysis. Microbiome. 2014;2:15.PubMed 
PubMed Central 
Article 

Google Scholar 
Gloor GB, Macklaim JM, Pawlowsky-Glahn V, Egozcue JJ. Microbiome datasets are compositional: and this is not optional. Front Microbiol. 2017;8:2224.PubMed 
PubMed Central 
Article 

Google Scholar 
Luo W, Friedman MS, Shedden K, Hankenson KD, Woolf PJ. GAGE: generally applicable gene set enrichment for pathway analysis. BMC Bioinformatics. 2009;10:161.PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Luo W, Brouwer C. Pathview: an R/Bioconductor package for pathway-based data integration and visualization. Bioinformatics. 2013;29:1830–1.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Rohart F, Gautier B, Singh A, Cao K-AL. mixOmics: An R package for ‘omics feature selection and multiple data integration. PLOS Computational Biology. 2017;13:e1005752.PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Palarea-Albaladejo J, Martín-Fernández JA. zCompositions—R package for multivariate imputation of left-censored data under a compositional approach. Chemometrics Intell Lab Syst. 2015;143:85–96.CAS 
Article 

Google Scholar 
Plaza Oñate F, Le Chatelier E, Almeida M, Cervino ACL, Gauthier F, Magoulès F, et al. MSPminer: abundance-based reconstitution of microbial pan-genomes from shotgun metagenomic data. Bioinformatics. 2019;35:1544–52.PubMed 
Article 
CAS 

Google Scholar 
Csardi G, Nepusz T. The Igraph Software Package for Complex Network Research. InterJournal 2006, Complex Systems, 1695.Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 2015;25:1043–55.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Epskamp S, Cramer AOJ, Waldorp LJ, Schmittmann VD, Borsboom D. qgraph: Network Visualizations of Relationships in Psychometric Data. J Stat Softw. 2012;48:1–18.Article 

Google Scholar 
Lambert S, Tragin M, Lozano J-C, Ghiglione J-F, Vaulot D, Bouget F-Y, et al. Rhythmicity of coastal marine picoeukaryotes, bacteria and archaea despite irregular environmental perturbations. ISME J. 2019;13:388–401.PubMed 
Article 

Google Scholar 
Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016;13:581–3.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Galand PE, Casamayor EO, Kirchman DL, Lovejoy C. Ecology of the rare microbial biosphere of the Arctic Ocean. PNAS. 2009;106:22427–32.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Campbell BJ, Yu L, Heidelberg JF, Kirchman DL. Activity of abundant and rare bacteria in a Coastal Ocean. Proc Natl Acad Sci USA. 2011;108:12776–81.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Morris RM, Rappé MS, Connon SA, Vergin KL, Siebold WA, Carlson CA, et al. SAR11 clade dominates ocean surface bacterioplankton communities. Nature. 2002;420:806–10.CAS 
PubMed 
Article 

Google Scholar 
Bouvier T, del Giorgio PA. Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol. 2007;9:287–97.CAS 
PubMed 
Article 

Google Scholar 
Thingstad TF, Våge S, Storesund JE, Sandaa R-A, Giske J. A theoretical analysis of how strain-specific viruses can control microbial species diversity. Proc Natl Acad Sci USA. 2014;111:7813–8.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Pedrós-Alió C. Marine microbial diversity: can it be determined? Trends Microbiol. 2006;14:257–63.PubMed 
Article 
CAS 

Google Scholar 
Gobet A, Böer SI, Huse SM, van Beusekom JEE, Quince C, Sogin ML, et al. Diversity and dynamics of rare and of resident bacterial populations in coastal sands. ISME J. 2012;6:542–53.PubMed 
Article 

Google Scholar 
Pascoal F, Costa R, Assmy P, Duarte P, Magalhães C. Exploration of the types of rarity in the arctic ocean from the perspective of multiple methodologies. Microb Ecol. 2021;84:59–72.PubMed 
Article 
CAS 

Google Scholar 
Huete-Stauffer TM, Arandia-Gorostidi N, Díaz-Pérez L, Morán XAG. Temperature dependences of growth rates and carrying capacities of marine bacteria depart from metabolic theoretical predictions. FEMS Microbiol Ecol. 2015;91:fiv111.PubMed 
Article 
CAS 

Google Scholar 
Arandia-Gorostidi N, Huete-Stauffer TM, Alonso-Sáez L, G. Morán XA. Testing the metabolic theory of ecology with marine bacteria: different temperature sensitivity of major phylogenetic groups during the spring phytoplankton bloom. Environ Microbiol. 2017;19:4493–505.CAS 
PubMed 
Article 

Google Scholar 
Giovannoni SJ, Bibbs L, Cho J-C, Stapels MD, Desiderio R, Vergin KL, et al. Proteorhodopsin in the ubiquitous marine bacterium SAR11. Nature. 2005;438:82–85.CAS 
PubMed 
Article 

Google Scholar 
Yilmaz P, Yarza P, Rapp JZ, Glöckner FO. Expanding the world of marine bacterial and archaeal clades. Front Microbiol. 2016;6:1524.PubMed 
PubMed Central 
Article 

Google Scholar 
Pedler BE, Aluwihare LI, Azam F. Single bacterial strain capable of significant contribution to carbon cycling in the surface ocean. Proc Natl Acad Sci USA. 2014;111:7202–7.CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Pereira O, Hochart C, Boeuf D, Auguet JC, Debroas D, Galand PE. Seasonality of archaeal proteorhodopsin and associated Marine Group IIb ecotypes (Ca. Poseidoniales) in the North Western Mediterranean Sea. ISME J. 2020;15:1302–16.PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Iverson V, Morris RM, Frazar CD, Berthiaume CT, Morales RL, Armbrust EV. Untangling Genomes from Metagenomes: Revealing an Uncultured Class of Marine Euryarchaeota. Science. 2012;335:587–90.CAS 
PubMed 
Article 

Google Scholar 
Pereira O, Hochart C, Auguet JC, Debroas D, Galand PE. Genomic ecology of Marine Group II, the most common marine planktonic Archaea across the surface ocean. MicrobiologyOpen. 2019;8:e00852.PubMed 
PubMed Central 

Google Scholar 
Tully BJ. Metabolic diversity within the globally abundant Marine Group II Euryarchaea offers insight into ecological patterns. Nat Commun. 2019;10:271.PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Xie W, Luo H, Murugapiran SK, Dodsworth JA, Chen S, Sun Y, et al. Localized high abundance of Marine Group II archaea in the subtropical Pearl River Estuary: implications for their niche adaptation. Environ Microbiol. 2018;20:734–54.CAS 
PubMed 
Article 

Google Scholar 
Jousset A, Bienhold C, Chatzinotas A, Gallien L, Gobet A, Kurm V, et al. Where less may be more: how the rare biosphere pulls ecosystems strings. ISME J. 2017;11:853–62.PubMed 
PubMed Central 
Article 

Google Scholar 
Bernard G, Pathmanathan JS, Lannes R, Lopez P, Bapteste E. Microbial dark matter investigations: how microbial studies transform biological knowledge and empirically sketch a logic of scientific discovery. Genome Biol Evol. 2018;10:707–15.PubMed 
PubMed Central 
Article 

Google Scholar 
Carradec Q, Pelletier E, Da Silva C, Alberti A, Seeleuthner Y, Blanc-Mathieu R, et al. A global ocean atlas of eukaryotic genes. Nature Communications. 2018;9:373.PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Thomas AM, Segata N. Multiple levels of the unknown in microbiome research. BMC Biology. 2019;17:48.PubMed 
PubMed Central 
Article 

Google Scholar  More

Hear the latest from the world of science, with Nick Petrić Howe and Benjamin Thompson.

Your browser does not support the audio element.

Download MP3

In this episode:00:46 Inequalities in US faculty hiringIn the US, where a person gained their PhD can have an outsized influence on their future career. Now, using a decade worth of data, researchers have shown there are stark inequalities in the hiring process, with 80% of US faculty trained at just 20% of institutions.Research article: Wapman et al.09:01 Research HighlightsHow wildlife can influence chocolate production, and the large planets captured by huge stars.Research Highlight: A chocoholic’s best friends are the birds and the batsResearch Highlight: Giant stars turn to theft to snag jumbo planets11:42 Briefing ChatWe discuss some highlights from the Nature Briefing. This time, what science says about grieving for a public figure, and why suburban Australians are sharing increasingly sophisticated measures to prevent cockatoos from opening wheelie bins.Nature News: Millions are mourning the Queen — what’s the science behind public grief?The Guardian: ‘Interspecies innovation arms race’: cockatoos and humans at war over wheelie bin raidsSubscribe to Nature Briefing, an unmissable daily round-up of science news, opinion and analysis free in your inbox every weekday.Never miss an episode: Subscribe to the Nature Podcast on Apple Podcasts, Google Podcasts, Spotify or your favourite podcast app. Head here for the Nature Podcast RSS feed. More

The theoretical modelThe model presented here builds on a recently developed metabolic theory based on biochemical kinetics. It describes a non-linear relationship between plant metabolic rate per unit of dry mass (Bs, nmol g−1 s−1), such as light-saturated photosynthetic rates and dark respiration rates, plant water content (S, g g−1), and temperature (in degrees K)19,20, i.e.$${B}_{{{{{{rm{s}}}}}}}={g}_{1}{e}^{{k}_{1}S/left({K}_{1}+Sright)}{e}^{-E/{kT}}$$
(1)
where g1 is a normalisation constant, k1 represents the maximum increase in specific metabolic rates due to changes in water content (i.e. from dehydrated to fully hydrated), K1 represents the water content when the mean reaction rate of cellular metabolism reaches one-half of its maximum, E is the activation energy, and k is Boltzmann’s constant. In this model S is defined on a dry mass basis (i.e. the ratio of plant water mass to plant dry mass) to broaden its range and better reflect the proportional changes in the amount of water in plant tissues20. Full details of the model’s assumptions can be found in the Methods and Huang et al.19,20. The model was tested and shown to hold true for a broad range of species and for whole plants and above- and belowground organs20. Here, we first applied the model to describe the quantitative effects of dry mass-based LWC (the ratio of leaf water mass to leaf dry mass) and temperature on the light-saturated leaf photosynthetic rate per unit of dry mass or leaf photosynthetic capacity (Ps, nmol CO2 g−1 s−1), which can be expressed as$${{{{{rm{ln}}}}}}left({P}_{{{mbox{s}}}}right)={{{{{rm{ln}}}}}}left(frac{{P}_{{{{{{rm{L}}}}}}}}{{M}_{{{{{{rm{L}}}}}}}}right)={{{{{rm{ln}}}}}}left({g}_{1}right)+frac{{k}_{1}cdot {{{{{rm{LWC}}}}}}}{{K}_{1}+{{{{{rm{LWC}}}}}}}-frac{E}{{kT}},$$
(2)
where PL is the light-saturated whole-leaf photosynthetic rate (nmol s−1). Equation 2 indicates that the log-transformed temperature-corrected Ps (i.e. Pscor) should increase with LWC, following Michaelis-Menten type hyperbolic response, i.e.$${{{{{rm{ln}}}}}}left({P}_{{{mbox{scor}}}}right)={{{{{rm{ln}}}}}}left({P}_{{{{{{rm{s}}}}}}}{e}^{E/{kT}}right)={{{{{rm{ln}}}}}}left({g}_{1}right)+frac{{k}_{1}cdot {{{{{rm{LWC}}}}}}}{{K}_{1}+{{{{{rm{LWC}}}}}}}.$$
(3)
Rearranging Eq. (2) shows that the temperature- and LWC-corrected whole-leaf photosynthetic rate (i.e. PLcor) should scale isometrically with ML, i.e.$${P}_{{{mbox{Lcor}}}}={P}_{{{{{{rm{L}}}}}}}{e}^{-{k}_{1}cdot {{{{{rm{LWC}}}}}}/left({K}_{1}+{{{{{rm{LWC}}}}}}right)}{e}^{E/{kT}}={g}_{1}{M}_{{{{{{rm{L}}}}}}}.$$
(4)
In this study, we refer to the temperature or water content correction as moving the temperature term ((frac{E}{{kT}})) or water content term ((frac{{k}_{1}cdot {{{{{rm{LWC}}}}}}}{{K}_{1}+{{{{{rm{LWC}}}}}}})) to the left-hand side of the model, an approach that has been commonly used in previous studies28.Following previous studies29,30,31, we assume that PL is proportional to AL, i.e. ({P}_{L}propto {A}_{L})because leaf area directly determines the light interception capacity1. Therefore, the quantitative relationship between SLA and LWC and temperature can be described as$${{{{{{mathrm{ln}}}}}}}left({{mbox{SLA}}}right)={{{{{{mathrm{ln}}}}}}}left(frac{{A}_{{{{{{rm{L}}}}}}}}{{M}_{{{{{{rm{L}}}}}}}}right)={{{{{{mathrm{ln}}}}}}}left({g}_{2}right)+frac{{k}_{1}cdot {{{{{rm{LWC}}}}}}}{{K}_{1}+{{{{{rm{LWC}}}}}}}-frac{E}{{kT}},$$
(5)
where g2 is another normalisation constant. Given that leaf area is a direct indicator of leaf photosynthetic capacity and that both traits reflect the long-term adaptation of plants to environmental change3,13, k1 in Eqs. (2) and (5) represent the maximum increase in mass-specific leaf photosynthetic capacity due to changes in water content. Since the temperature has a direct effect on the metabolic rates32 and productivity of ecosystems33, it is reasonable to assume that temperature can affect SLA globally34,35. Therefore, in this context T denotes the mean growing-season temperature (in degrees K), as SLA might be more responsive to long-term changes in temperature. Equation (5) predicts that SLA should increase with both LWC and the growing-season temperature. Rearranging Eq. (5) yields$${{{{{{mathrm{ln}}}}}}}left({{mbox{SL}}}{{{mbox{A}}}}_{{{mbox{cor}}}}right)={{{{{{mathrm{ln}}}}}}}left({{mbox{SLA}}}{e}^{E/{kT}}right)={{{{{{mathrm{ln}}}}}}}left({g}_{2}right)+frac{{k}_{1}cdot {{{{{rm{LWC}}}}}}}{{K}_{1}+{{{{{rm{LWC}}}}}}},$$
(6)
which predicts that the log-transformed temperature-corrected SLA (i.e. SLAcor) should increase with LWC following Michaelis-Menten dynamics. Likewise, by moving (frac{{k}_{1}cdot {{{{{rm{LWC}}}}}}}{{K}_{1}+{{{{{rm{LWC}}}}}}}) and (frac{E}{{kT}}) to the left-hand side and ML to the right-hand side of Eq. (5), we observe that the temperature- and LWC-corrected leaf area (i.e. ALcor) should scale isometrically with leaf mass, i.e.$${A}_{{{mbox{Lcor}}}}={A}_{{{{{{rm{L}}}}}}}{e}^{-{k}_{1}cdot {{{{{rm{LWC}}}}}}/left({K}_{1}+{{{{{rm{LWC}}}}}}right)}{e}^{E/{kT}}={g}_{2}{M}_{{{{{{rm{L}}}}}}}.$$
(7)
We note that the scaling of PLcor and ALcor with respect to ML will reveal how LWC mediates the scaling exponent of leaf trait relationships.Effects of LWC and temperature on leaf trait scalingThe numerical value of the exponent for the PL versus AL scaling relationship calculated from the empirical data was 0.99 (Supplementary Fig. 1; 95% CI = 0.95 and 1.02, r2 = 0.87), strongly supporting the model assumption that PL scales isometrically with AL. We then examined the effect of LWC on leaf trait scaling. The numerical value of the scaling exponent for the PL versus ML relationship was 0.95 (Fig. 2a; 95% CI = 0.92 and 0.99, r2 = 0.83). The non-linear relationship between Pscor and LWC described by Eq. (3) was supported by the empirical data (Fig. 2b; Supplementary Table 1). The non-linear model (Eq. 3) also had a lower Akaike’s Information Criterion score than the simple linear model between log-transformed Pscor and LWC (i.e. 793.5 versus 1988.8). After LWC and temperature were corrected (see Eq. 4), the numerical value of the scaling exponent became 0.97 (Fig. 2c; 95% CI = 0.94 and 1.01, r2 = 0.85), which was statistically indistinguishable from 1.0 (P  > 0.05), as predicted by the model. Likewise, the numerical value of the exponent (i.e. α) for the AL versus ML scaling relationship was 1.02 (Fig. 2d; 95% CI = 1.02 and 1.03, r2 = 0.92). Additional analyses using the pooled dataset showed that log-transformed SLA increased with LWC following Michaelis-Menten dynamics, as predicted by Eq. (6) (Fig. 2e; Supplementary Table 1). After the effects of LWC and temperature were accounted for (using Eq. 7), the numerical value of α became 1.01 (Fig. 2f; 95% CI = 1.00 and 1.01, r2 = 0.95). Thus, both of the scaling exponents numerically converged onto 1.0 once LWC and temperature were corrected. In addition, an inspection of the locally weighted smoothing (LOWESS) curves showed that the curvature in both scaling relationships was reduced after the effects of LWC and temperature were corrected, as predicted by the model (Fig. 2).Fig. 2: The quantitative effects of dry mass-based leaf water content (LWC) on leaf photosynthesis and SLA.a Scaling of leaf photosynthetic rate (PL, nmol s−1) with leaf dry mass (ML, g). b Non-linear fit to the relationship between temperature-corrected mass-specific leaf photosynthetic rate (Pscor, nmol g−1 s−1) and LWC (g g−1) based on Eq. (3). c Scaling of temperature- and LWC-corrected leaf photosynthetic rate (PLcor, nmol s−1) with ML (g). d Scaling of leaf area (AL, cm2) with leaf dry mass (ML, g). e Non-linear fit to the relationship between temperature-corrected specific leaf area (SLAcor, cm2 g−1) and LWC based on Eq. (6). f Scaling of temperature- and LWC-corrected leaf area (ALcor, cm2) with leaf dry mass (ML, g). Data with LWC greater than 25 were not shown in panel e for a better visualisation. LOWESS curves (blue lines) and 95% confidence intervals are shown.Full size imageThe results presented here show that temperature and LWC quantitatively correlate with other leaf traits, such as Ps and SLA (or LMA), as predicted by the model. The increases in Ps and SLA attenuate with increasing LWC (Fig. 2b, e), indicating that leaf water availability sets a constraint on the maximum Ps and SLA that leaves can reach. It has long been recognised that SLA is closely correlated with leaf growth rate and metabolic activity14,36,37. Therefore, it is reasonable to also expect that SLA, as well as Ps, will be quantitatively affected by LWC, which can change as a function of developmental status (such as leaf maturation and the accumulation of lignified tissues) and transiently as a function of evapotranspiration. LWC is also a reflection of species-specific adaptation to environmental conditions in different biomes. Nevertheless, our model, as well as the empirical data used to test it, reveal a broad and statistically robust correlation between critical leaf functional traits and leaf tissue water content. However, the observed variations in Pscor (Fig. 2b) suggest that in addition to temperature and LWC, other factors (e.g. plant phylogeny and soil fertility) may also affect leaf photosynthetic capacity, which is not accounted for in our model and should be critically examined in future research.Leaf area-mass scaling among different groupsThe numerical value of α varied across different plant growth forms, ecosystems, and latitudinal zones (Fig. 3 and Supplementary Table 2). In particular, the leaf area versus mass scaling relationship showed a clear pattern along a latitudinal gradient (Fig. 3c and Supplementary Table 2). The numerical value of α decreased from 1.10 in boreal regions (95% CI = 1.08 and 1.12, r2 = 0.91) to 1.00 in temperate regions (95% CI = 0.99 and 1.01, r2 = 0.91), and to 0.94 in tropical regions (95% CI = 0.92 and 0.95, r2 = 0.91). However, after correcting for the effects of LWC and temperature, as predicted, α converged onto 1.0 across all different groupings, and the r2 values of the scaling relationships also increased (Fig. 3 and Supplementary Table 2).Fig. 3: The exponents of leaf area-mass scaling with and without temperature and LWC corrections among different groups.a Comparison of scaling exponents among plant growth forms (n = 1688, 491, 1097, and 832 for forbs, graminoids, shrubs, and trees, respectively). b Comparison of scaling exponents among ecosystem types (n = 97, 1367, 1285, 1271, and 114 for deserts, forests, grasslands, tundra, and wetlands, respectively). c Comparison of scaling exponents among different latitudinal zones (n = 1111, 2113, and 910 for tropical, temperate, and boreal zones, respectively). Error bars indicate 95% confidence intervals.Full size imageOur analyses show that LWC also affects the numerical values of the exponents of leaf trait scaling relationships, which helps to explain why different works sometimes report significant differences in the exponents governing these relationships8,10. In particular, the numerical values of the scaling exponent governing the leaf area versus mass scaling relationship differ among different plant growth forms, ecosystems, and latitudinal zones (Fig. 3 and Supplementary Table 2), indicating that no invariant “scaling exponent” (i.e. α) holds true for the leaf area-mass scaling relationship. For example, in our dataset, α is significantly smaller than 1.0 in tropical regions (i.e. in keeping with a “diminishing returns” relationship in leaf area with respect to increasing leaf mass), close to 1.0 in temperate regions (i.e. a break-even relationship), and significantly larger than 1.0 in boreal regions (i.e. an “increasing returns” relationship). This shift in α along a latitudinal gradient may be associated with the different strategies to cope with variations in water availability, e.g. the high evapotranspiration rates in tropical regions may constrain increases in leaf area with increasing leaf mass, therefore resulting in diminishing returns, whereas, in boreal regions, the reduced water stress may enable plants to maximise leaf area to achieve relatively high photosynthetic capacities. Despite the variability in the numerical values of scaling exponents across different groups, the degree of curvature in these scaling relationships is reduced, and exponents converge onto unity after the effects of LWC and temperature are accounted for (Fig. 3), as predicted by the model (Eq. 7). This finding indicates that the variations in the exponent can, at least partially, be ascribed to the effects of LWC on SLA and the relative rate of increase in leaf area versus leaf mass. It is noteworthy that the numerical value of the scaling exponent for the PL versus ML relationship is also very close to 1.0, indicating the relatively weak effects of temperature and LWC on leaf photosynthesis-mass scaling. This may be partially attributed to the relatively limited number of data with concurrent LWC measurements and the adaptation of leaf traits to long-term temperature changes (see below for detailed discussion). Nevertheless, more measurements on LWC and leaf photosynthetic rates are needed to further test how LWC mediates leaf photosynthesis-mass scaling.Given that LWC was weakly correlated with ML (Supplementary Fig. 2, log-log slope = −0.01, 95% CI = −0.02 and 0, r2  More

Leston, D. & Leston, D. The ant mosaic-tropical tree crops and the limiting of pests and diseases. PANS Pest Artic. News Summ. 19, 311–341 (1973).Article 

Google Scholar 
Huang, H. T. & Yang, P. A tropical ant is used to control insect pests in southern China. Bioscience 37, 665–671 (1987).Article 

Google Scholar 
Peng, R. K. & Christian, K. The weaver ant, Oecophylla smaragdina (Hymenoptera: Formicidae), an effective biological control agent of the red-banded thrips, Selenothrips rubrocinctus (Thysanoptera: Thripidae) in mango crops in the Northern Territory of Australia. Int. J. Pest Manag. 50, 107–114 (2004).Article 

Google Scholar 
Peng, R. K. & Christian, K. The control efficacy of the weaver ant, Oecophylla smaragdina (Hymenoptera: Formicidae), on the mango leafhopper, Idioscopus nitidulus (Hemiptera: Cicadellidea) in mango orchards in the northern territory). Int. J. Pest Manag. 51, 297–304 (2005).Article 

Google Scholar 
Peng, R. K. & Christian, K. Effective control of Jarvis’s fruit fly, Bactrocera jarvisi (Diptera: Tephritidae), by the weaver ant, Oecophylla smaragdina (Hymenoptera: Formicidae), in mango orchards in the Northern Territory of Australia. Int. J. Pest Manag. 52, 275–282 (2006).Article 

Google Scholar 
Peng, R. K. & Christian, K. The effect of the weaver ant, Oecophylla smaragdina (Hymenoptera: Formicidae), on the mango seed weevil, Sternochetus mangiferae (Coleoptera: Curculionidae), in mango orchards in the Northern Territory of Australia. Int. J. Pest Manag. 53, 15–24 (2007).Article 

Google Scholar 
Blüthgen, N. & Stork, N. E. Ant mosaics in a tropical rainforest in Australia and elsewhere: A critical review. Aust. Ecol. 32, 93–104 (2007).Article 

Google Scholar 
Davidson, D. W., Lessard, J. P., Bernau, C. R. & Cook, S. C. The tropical ant mosaic in a primary Bornean rain forest. Biotropica 39, 468–475 (2007).Article 

Google Scholar 
Peng, R. K., Christian, K. & Gibb, K. The effect of colony isolation of the predacious ant, Oecophylla smaragdina (f.) (Hymenoptera: Formicidae), on protection of cashew plantations from insect pests. Int. J. Pest Manag. 45, 189–194 (1999).Article 

Google Scholar 
Peng, R. K., Christian, K. & Gibb, K. Ecology of the fruit spotting bug, Amblypelta lutescens lutescens Distant (Hemiptera: Coreidae) in cashew plantations, with particular reference to the potential for its biological control. Aust. J. Entomol. 44, 45–51 (2005).Article 

Google Scholar 
Van Mele, P. & Cuc, N. T. T. Evolution and status of Oecophylla smaragdina (Fabricius) as a pest control agent in citrus in the Mekong Delta, Vietnam. Int. J. Pest Manag. 46, 295–301 (2000).Article 

Google Scholar 
Van Mele, P., Cuc, N. T. T. & VanHuis, A. Direct and indirect influences of the weaver ant Oecophylla smaragdina on citrus farmers’ pest perceptions and management practices in the Mekong Delta, Vietnam. Int. J. Pest Manag. 48, 225–232 (2002).Article 

Google Scholar 
Kumaresan, V. Prevention of rhinoceros beetle (Oryctes rhinoceros) in coconut palm using red ants. J. Bombay Nat. Hist. Soc. 93, 308–309 (1996).
Google Scholar 
Way, M. J. & Khoo, K. C. Relationships between Helopeltis theobromae damage and ants with special reference to Malaysian cocoa small holdings. J. Plant Prot. Trop. 6, 1–11 (1989).
Google Scholar 
Way, M. J. & Khoo, K. C. Colony dispersion and nesting habits of the ants, Dolichoderus thoracicus and Oecophylla smaragdina (Hymenoptera: Formicidae), in relation to their success as biological control agents on cocoa. Bull. Entomol. Res. 81, 341–350 (1991).Article 

Google Scholar 
Van Mele, P. Biological Control with the Weaver Ant, Oecophylla longinoda, in Africa: A Review of Research and Development Efforts to Link Farmers to Organic Markets. Outlooks Pest Manag. 19, 180–183 (2008).Article 

Google Scholar 
Van Mele, P., Vayssieres, J. F., Tellingen, E. V. & Vrolijks, J. Effects of an African weaver ant, Oecophylla longinoda, in controlling mango fruit flies (Diptera: Tephritidae) in Benin. J. Econ. Entomol. 100, 695–701 (2007).PubMed 
Article 

Google Scholar 
Van Mele, P., Vayssieres, J. F., Adandonon, A. & Sinzogan, A. Ant cues affect the oviposition behaviour of fruit flies (Diptera: Tephritidae) in Africa. Physiol. Entomol. 34, 256–261 (2009).Article 

Google Scholar 
Kempraj, V., Park, S. P. & Taylor, P. W. Forewarned is forearmed: Queensland fruit flies detect olfactory cues from predators and respond with predator-specific behaviour. Sci. Rep. 10, 7297 (2020).ADS 
CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Kempraj, V., Park, S. J. & Faveri, D. S. Overlooked scents: Chemical profile of soma, volatile emissions and trails of the green tree ant, Oecophylla smaragdina. Molecules 25, 2112 (2020).CAS 
PubMed Central 
Article 

Google Scholar 
Burks, R. L. & Lodge, D. M. Cued in: Advances and opportunities in freshwater chemical ecology. J. Chem. Ecol. 28, 1901–1917 (2002).CAS 
PubMed 
Article 

Google Scholar 
Fink, P. Ecological functions of volatile organic compounds in aquatic systems. Mar. Freshw. Behav. Physiol. 40, 155–168 (2007).CAS 
Article 

Google Scholar 
Spencer, M., Blaustein, L. & Cohen, J. E. Oviposition habitat selection by mosquitoes (Culiseta longiareolata) and consequences for population size. Ecology 83, 669–679 (2002).Article 

Google Scholar 
Binckley, C. A. & Resetarits, W. J. Habitat selection determines abundance, richness and species composition of beetles in aquatic communities. Biol. Lett. 1, 370–374 (2005).PubMed 
PubMed Central 
Article 

Google Scholar 
Biswas, M. J. H., Mainali, B., Park, S. J., Taylor, P. & Rempoulakis, P. Electrophysiological responses to cuelure of raspberry ketone-fed Queensland fruit fly. J. Econ. Entomol. 113, 2832–2839 (2020).CAS 
PubMed 
Article 

Google Scholar 
Kats, L. B. & Dill, L. M. The scent of death: Chemosensory assessment of predation risk by prey animals. Ecoscience 5, 361–394 (1998).Article 

Google Scholar 
Brönmark, C. & Hansson, L. A. Chemical communication in aquatic systems: An introduction. Oikos 88, 103–109 (2000).Article 

Google Scholar 
Dicke, M. & Grostal, P. Chemical detection of natural enemies by arthropods: An ecological perspective. Annu. Rev. Ecol. Syst. 32, 1–23 (2001).Article 

Google Scholar 
Van Donk, E. Chemical information transfer in freshwater plankton. Ecol. Inform. 2, 112–120 (2007).Article 

Google Scholar 
Pohnert, G., Steinke, M. & Tollrian, R. Chemical cues, defence metabolites and the shaping of pelagic interspecific interactions. Trends Ecol. Evol. 22, 198–204 (2007).PubMed 
Article 

Google Scholar 
Tollrian, R. & von Elert, E. Enrichment and purification of Chaoborus kairomone from water: Further steps toward its chemical characterization. Limnol. Oceanogr. 39, 788–796 (1994).ADS 
CAS 
Article 

Google Scholar 
Von Elert, E. & Loose, C. J. Predator-induced diel vertical migration in Daphnia: Enrichment and preliminary chemical characterization of a kairomone exuded by fish. J. Chem. Ecol. 22, 885–895 (1996).Article 

Google Scholar 
Kusch, J. Self-recognition as the original function of an amoeban defense-inducing kairomone. Ecology 80, 715–720 (1999).Article 

Google Scholar 
Peerzada, N., Pakkiyaretnam, T. & Renaud, S. Volatile constituents of the Green Ant Oecophylla smaragdina. Agric. Biol. Chem. 54, 3335–3336 (1990).CAS 

Google Scholar 
Keegans, S. J., Billen, J. & Morgan, E. D. Volatile secretions of the green tree ant Oecophylla smaragdina (Hymenoptera: Formicidae). Comp. Biochem. Physiol. Part B Biochem. 100, 681–685 (1991).Article 

Google Scholar 
Kraus, B. effects of honey-bee alarm pheromone compounds on the behaviour of Varroa jacobsoni. Apidologie 21, 127–134 (1990).CAS 
Article 

Google Scholar 
Brossut, R. Allomonal secretions in cockroaches. J. Chem. Ecol. 9, 143–158 (1983).CAS 
PubMed 
Article 

Google Scholar 
Ali, A. et al. Insecticidal and biting deterrent activities of Magnolia grandiflora essential oil and selected pure compounds against Aedes aegypti. Molecules 25, 1359 (2020).CAS 
PubMed Central 
Article 

Google Scholar 
Yu, J., Yang, B., Chang, Y., Zhang, Y. & Wang, G. Identification of a general odorant receptor for repellents in the Asian corn borer Ostrinia furnacalis. Front. Physiol. 11, 1–43. https://doi.org/10.3389/fphys.2020.00176 (2020).CAS 
Article 

Google Scholar 
Sievert, T. et al. Bank vole alarm pheromone chemistry and effects in the field. Oecologia 196, 667–677 (2021).ADS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Fletcher, B. S. & Watson, C. A. The ovipositional response of the tephritid fruit fly, Dacus tryoni to 2-Chloroethanol in laboratory bioassays. Ann. Entomol. Soc. Am. 67, 21–23 (1974).CAS 
Article 

Google Scholar 
Eisemann, C. H. & Rice, M. J. Attractants for the gravid Queensland fruit fly Dacus tryoni. Entomol. Exp. Appl. 62, 125–130 (1992).CAS 
Article 

Google Scholar 
Devescovi, F., Hurtado, J. & Taylor, W. P. Mating-induced changes in responses of female Queensland fruit fly to male pheromones and fruit: A mechanism for mating-induced sexual inhibition. J. Insect Physiol. 129, 104195 (2021).CAS 
PubMed 
Article 

Google Scholar 
Kempraj, V., Park, S. P. & Taylor, P. W. γ-Octalactone, an effective oviposition stimulant of Bactrocera tryoni. J. Appl. Entomol. 143, 1205–1209 (2019).CAS 
Article 

Google Scholar 
Jayanthi, P. D. K. et al. Oviposition site-selection by Bactrocera dorsalis is mediated through an innate recognition template tuned to γ-octalactone. PLoS ONE 9, e85764 (2014).ADS 
Article 
CAS 

Google Scholar 
Weldon, C. V. & Taylor, P. W. Sexual development of wild and mass-reared male Queensland fruit fly in response to natural food sources. Entomol. Exp. Appl. 139, 17–24 (2011).Article 

Google Scholar 
Friard, O. & Gamba, M. BORIS: a free, versatile open-source event-logging software for video/audio coding and live observations. Methods Ecol. Evol. 7, 1325–1330 (2016).Article 

Google Scholar  More

Preliminary experiment using grapevine shoot samplesTo determine whether the profiles of endophytic bacterial microbiota vary widely between shoot samples collected from the same grapevine plant or between shoot samples collected from different grapevine plants of the same cultivar grown in the same vineyard, a preliminary experiment was performed. Microbiome analysis demonstrated that the profiles of endophytic bacterial microbiota were similar between two shoot samples collected from the same Chardonnay or Koshu grapevine plant (Fig. S2). In addition, the profiles of endophytic bacterial microbiota in shoot samples collected from different Chardonnay or Koshu grapevine plants cultivated in the same vineyard were also similar (Fig. S2). These results suggest that the profiles of endophytic bacterial microbiota in shoot samples collected from different grapevine plants of the same cultivar grown in the same vineyard were uniform. On the basis of this finding, we collected one shoot sample from a grapevine plant, at two different shoot growth stages (shoot elongation stage and véraison), of each cultivar grown in the eleven vineyards located in major grapevine-growing regions in Japan.Weather dataGDDs from April 1 to October 31, 2020 demonstrated that Minamisanriku and Ueda belonged to Region III on the Winkler Index and that Komoro, Shobara, and Saijo belonged to Region IV on the Winkler Index (Supplementary Table 2). Only Urausu belonged to Region II on the Winkler Index. Five vineyards including Kofu, Kai, Katsunuma, Izumo, and Omishima belonged to Region V on the Winkler Index, suggesting that Chardonnay, Pinot Noir, and Cabernet Sauvignon were cultivated under extremely high temperatures in those vineyards. Precipitation from April 1 to October 31, 2020 exceeded 1700 mm in Shobara, the highest among the vineyards (Supplementary Table 2).Amplicon sequences collected from grapevine shoot xylemsA total of 7,019,600 amplicon sequences from 52 samples were collected (Supplementary Table 3). We identified a total of 1305 OTUs on the basis of the conventional criterion of 99% sequence similarity. Irrespective of cultivar, grapevine-growing region, and shoot growth stage, Alphaproteobacteria, Gammaproteobacteria, and Oxyphotobacteria were predominant in shoot xylems (Fig. 1). Actinobacteria, Bacteroidia, Bacilli, and Clostridia were the endophytic bacteria detected in the shoot xylems.Figure 1Endophytic bacterial microbiota in shoot xylems of cultivars grown in the same vineyard. Endophytic bacterial microbiota in the shoot xylems of each cultivar collected from nine vineyards were identified and evaluated at the class level. Data are presented as relative abundance (%). KO Koshu, CH Chardonnay, CS Cabernet Sauvignon, PN Pinot Noir, SES shoot elongation stage, V véraison.Full size imageComparison of endophytic bacterial microbiota in grapevine shoot xylems of cultivars grown in the same vineyardShoot samples of two or more cultivars were collected from nine vineyards (Urausu, Katsunuma, Kofu, Kai, Komoro, Ueda, Izumo, Shobara, and Saijo) and evaluated (Fig. 1). Below are the detailed results for each vineyard.Urausu (Hokkaido Prefecture)At the shoot elongation stage, more than 90% of endophytic bacteria in Chardonnay and Pinot Noir shoot xylems belonged to class Gammaproteobacteria. Oxyphotobacteria was also detected in the shoot xylems albeit at a very low proportion (1% and 2% in Chardonnay and Pinot Noir, respectively). At véraison, the proportion of Oxyphotobacteria increased and reached 75% and 74% in Chardonnay and Pinot Noir shoot xylems, respectively. Overall, the profiles of endophytic bacterial microbiota were very similar between Chardonnay and Pinot Noir cultivated in Urausu at each shoot growth stage.Katsunuma (Yamanashi Prefecture)At the shoot elongation stage, Gammaproteobacteria was predominant in Koshu and Cabernet Sauvignon shoot xylems, although Oxyphotobacteria and Bacilli were detected as well. At véraison, the proportion of Oxyphotobacteria increased and reached 48% and 75% in Koshu and Cabernet Sauvignon shoot xylems, respectively. The proportion of Alphaproteobacteria also increased at véraison (37% and 15% in Koshu and Cabernet Sauvignon, respectively). Overall, the profiles of endophytic bacterial microbiota were similar between Koshu and Cabernet Sauvignon cultivated in Katsunuma at each shoot growth stage.Kofu (Yamanashi Prefecture)Shoot samples of Koshu, Chardonnay, Pinot Noir, and Cabernet Sauvignon were collected from Kofu. At the shoot elongation stage, Gammaproteobacteria was predominant (approximately 90%) in Koshu and Pinot Noir shoot xylems, whereas more than 80% of endophytic bacteria in Chardonnay and Cabernet Sauvignon shoot xylems belonged to class Oxyphotobacteria. At véraison, the profiles of endophytic bacterial microbiota were similar among the four cultivars grown in Kofu, and Oxyphotobacteria was predominant.Kai (Yamanashi Prefecture)Irrespective of the shoot growth stage, Gammaproteobacteria was predominant in Chardonnay and Cabernet Sauvignon shoot xylems. Although Gammaproteobacteria was also predominant in the Koshu shoot xylems at the shoot elongation stage, the proportions of Oxyphotobacteria and Alphaproteobacteria increased in Koshu shoot xylems at véraison (57% and 32%, respectively).Komoro (Nagano Prefecture)Irrespective of the cultivar (Chardonnay, Pinot Noir, and Cabernet Sauvignon), the profiles of endophytic bacterial microbiota in shoot xylems were very similar at each shoot growth stage, and Oxyphotobacteria was predominant. More than 80% of endophytic bacteria in the shoot xylems at véraison belonged to class Oxyphotobacteria.Ueda (Nagano Prefecture)The profiles of endophytic bacterial microbiota in shoot xylems at the shoot elongation stage were similar among Chardonnay, Pinot Noir, and Cabernet Sauvignon, whereas the profile in Chardonnay shoot xylems at véraison was different from those in Pinot Noir and Cabernet Sauvignon shoot xylems. Gammaproteobacteria (76%) was predominant in Chardonnay shoot xylem at véraison. In Pinot Noir and Cabernet Sauvignon shoot xylems at véraison, more than 70% of endophytic bacteria belonged to class Oxyphotobacteria.Izumo (Shimane Prefecture)Unlike other vineyards, there was no similarity of profiles between cultivars (Chardonnay and Cabernet Sauvignon) and between shoot growth stages. Gammaproteobacteria and Oxyphotobacteria were predominant in Chardonnay shoot xylems at the shoot elongation stage and véraison, respectively. In Cabernet Sauvignon shoot xylems, Gammaproteobacteria (36% and 52% at the shoot elongation stage and véraison, respectively) and Oxyphotobacteria (34% and 43% at the shoot elongation stage and véraison, respectively) were predominant irrespective of the shoot growth stage.Shobara (Hiroshima Prefecture)Similarly to Urausu and Katsunuma, Gammaproteobacteria was predominant in Chardonnay and Cabernet Sauvignon shoot xylems at the shoot elongation stage. The proportion of Oxyphotobacteria increased at véraison; more than 70% of endophytic bacteria in Chardonnay and Cabernet Sauvignon shoot xylems at véraison belonged to class Oxyphotobacteria. Overall, the profiles of endophytic bacterial microbiota were similar between Chardonnay and Cabernet Sauvignon cultivated in Shobara at each shoot growth stage.Saijo (Hiroshima Prefecture)Similarly to Urausu, Katsunuma, and Shobara, Gammaproteobacteria (89%, 89%, and 98% in Koshu, Pinot Noir, and Cabernet Sauvignon shoot xylems, respectively) was predominant at the shoot elongation stage and Oxyphotobacteria (60%, 56%, and 63% in Koshu, Pinot Noir, and Cabernet Sauvignon shoot xylems, respectively), at véraison. Overall, the profiles of endophytic bacterial microbiota were similar among Koshu, Pinot Noir, and Cabernet Sauvignon cultivated in Saijo at each shoot growth stage.Comparison of endophytic bacterial microbiota in grapevine shoot xylems of cultivars grown in different vineyardsThe profiles of endophytic bacterial microbiota in the shoot xylems of Koshu, Chardonnay, Pinot Noir, and Cabernet Sauvignon cultivated in different vineyards were evaluated (Fig. 2). In Koshu shoot xylems, the profiles of endophytic bacterial microbiota were similar at each shoot growth stage irrespective of the vineyard. Gammaproteobacteria (73–89%) was predominant in Koshu shoot xylems at the shoot elongation stage, whereas Oxyphotobacteria (48–63%) and Alphaproteobacteria (19–37%) were predominant at véraison. At the shoot elongation stage, Pinot Noir cultivated in Komoro showed different diversity of endophytic bacterial microbiota from Pinot Noir cultivated in the other vineyards. At véraison, the profiles of endophytic bacterial microbiota in Pinot Noir shoot xylems were similar irrespective of the vineyard. Gammaproteobacteria (76–98%) was predominant in Pinot Noir shoot xylems at the shoot elongation stage, whereas Oxyphotobacteria (56–81%) was predominant at véraison. In contrast to Koshu and Pinot Noir, the profiles of endophytic bacterial microbiota in Chardonnay and Cabernet Sauvignon shoot xylems showed diversity and complexity among vineyards. At the shoot elongation stage, Oxyphotobacteria was predominant in Chardonnay shoot xylems at Minamisanriku (70%) and Kofu (85%), whereas Gammaproteobacteria was predominant in the other vineyards. At véraison, more than 95% of endophytic bacteria in shoot xylems of Chardonnay cultivated in Minamisanriku and Omishima belonged to class Gammaproteobacteria. In the case of Cabernet Sauvignon, although Oxyphotobacteria and Gammaproteobacteria were predominant in shoot xylems at both shoot elongation stage and véraison, their proportions drastically varied among vineyards.Figure 2Endophytic bacterial microbiota in shoot xylems of cultivars grown in the different vineyards. Endophytic bacterial microbiota in the shoot xylems of each cultivar collected from different vineyards were identified and evaluated at the class level. Data are presented as relative abundance (%). KO Koshu, CH Chardonnay, CS Cabernet Sauvignon, PN Pinot Noir, SES shoot elongation stage, V véraison, UR Urausu, MS Minamisanriku, KF Kofu, KA Kai, KN Katsunuma, KM Komoro, UE Ueda, SH Shobara, IZ Izumo, SA Saijo, OM Omishima.Full size imageComparison of endophytic bacterial microbiota in grapevine shoot xylems between shoot elongation stage and véraisonThe profiles of endophytic bacterial microbiota in the shoot xylems, regardless of the cultivar, at each shoot growth stage were evaluated (Fig. 3). The profiles of endophytic bacterial microbiota in grapevine shoot xylems at the shoot elongation stage were diverse and complex. Although Oxyphotobacteria and Gammaproteobacteria were predominant in the shoot xylems at the shoot elongation stage, various endophytic bacteria including those belonging to classes Actinobacteria, Bacteroidia, Bacilli, Clostridia, and Alphaproteobacteria existed in the shoot xylems as well. In contrast, the profiles of endophytic bacterial microbiota in grapevine shoot xylems at véraison showed far less variation than those at the shoot elongation stage. Oxyphotobacteria, Alphaproteobacteria, and Gammaproteobacteria accounted for more than 95% of endophytic bacteria in the shoot xylems at véraison.Figure 3Endophytic bacterial microbiota in grapevine shoot xylems at shoot elongation stage and véraison. Endophytic bacterial microbiota in the shoot xylems collected at the shoot elongation stage and véraison were identified and evaluated at the class level. Data are presented as relative abundance (%). UR Urausu, MS Minamisanriku, KF Kofu, KA Kai, KN Katsunuma, KM Komoro, UE Ueda, SH Shobara, IZ Izumo, SA Saijo, OM Omishima, KO Koshu, CH Chardonnay, CS Cabernet Sauvignon, PN Pinot Noir.Full size imageAlpha diversity of endophytic bacterial microbiota in grapevine shoot xylemsOTUs, Chao1 index, and Shannon index were used as indexes of alpha diversity of endophytic bacterial microbiota among cultivars, shoot growth stages, and vineyards (Fig. 4). The medians of OTUs were similar among the four cultivars (60.5 for Koshu and Pinot Noir, and 62.5 for Chardonnay and Cabernet Sauvignon). The medians of the Chao1 index were also comparable among the four cultivars (60 for Koshu and Pinot Noir, 63 for Chardonnay, and 65 for Cabernet Sauvignon). The median of the Shannon index (2.8) was highest for Koshu, whereas those for Pinot Noir, Chardonnay, and Cabernet Sauvignon were similar (2.0, 1.9, and 2.1, respectively). These results suggest that Koshu shoot xylems had a higher diversity of endophytic bacterial microbiota than Pinot Noir, Chardonnay, and Cabernet Sauvignon shoot xylems.Figure 4Alpha diversity of endophytic bacterial microbiota in grapevine shoot xylems. Alpha diversity analyses of cultivars, shoot growth stages, and vineyards were performed. Upper panels, OTUs; middle panels, Chao1 index; lower panels, Shannon index. Cross (×) indicates the average for each sample. KO Koshu, CH Chardonnay, CS Cabernet Sauvignon, PN Pinot Noir, SES shoot elongation stage, V véraison, UR Urausu, MS Minamisanriku, KF Kofu, KA Kai, KN Katsunuma, KM Komoro, UE Ueda, SH Shobara, IZ Izumo, SA Saijo, OM Omishima.Full size imageThe medians of OTUs and Chao1 index at the shoot elongation stage were comparable to those at véraison. The median of the Shannon index at the shoot elongation stage (3.0) was higher than that at véraison (1.7), indicating that grapevine shoot xylems at the shoot elongation stage had a higher diversity of endophytic bacterial microbiota than those at véraison.The medians of OTUs and Chao1 index were the highest for Ueda (74.5 and 75, respectively), whereas those were the lowest for Komoro (49.5 and 50, respectively). The medians of the Shannon index were lowest and highest for Minamisanriku (1.2) and Ueda (3.9), respectively. These results suggest that a large number of endophytic bacterial species existed in the shoot xylems of grapevine cultivated in Ueda, and that Ueda had the highest diversity of endophytic bacterial microbiota among the vineyards tested.Beta diversity of endophytic bacterial microbiota in grapevine shoot xylemsPCoA demonstrated that the plots of Koshu and Pinot Noir were relatively close to each other at the shoot elongation stage and very close to each other at véraison irrespective of the vineyard (Fig. 5), suggesting that the profiles of endophytic bacterial microbiota in Koshu and Pinot Noir shoot xylems were similar irrespective of both shoot growth stage and vineyard. Although the plots of Chardonnay and Cabernet Sauvignon in each vineyard were widely scattered at the shoot elongation stage, they were very close to each other at véraison. These results suggest that the profiles of endophytic bacterial microbiota in grapevine shoot xylems at véraison were uniform irrespective of the vineyard.Figure 5Principal coordinate analysis of endophytic bacterial microbiota in grapevine shoot xylems. Circles (○) and squares (□) indicate endophytic bacterial microbiota at the shoot elongation stage and véraison, respectively. KO Koshu, CH Chardonnay, CS Cabernet Sauvignon, PN Pinot Noir.Full size imagePERMANOVA demonstrated that the p-values for all combinations of cultivars exceeded 0.05 (Supplementary Table 4). In contrast, there was a significant difference (p = 0.001) between the shoot elongation stage and véraison. Although three of fifty-five combinations of vineyards showed significant differences (p = 0.04 for Komoro and Izumo, p = 0.007 for Komoro and Kai, and p = 0.034 for Kai and Kofu), there was no significant difference between most of the combinations. These results suggest that the variations of endophytic bacterial microbiota in grapevine shoot xylems greatly depended on the shoot growth stage.Cluster analysis of endophytic bacterial microbiota in grapevine shoot xylemsCluster analysis of endophytic bacterial microbiota in grapevine shoot xylems in various cultivars, shoot growth stages, and vineyards was performed by MDS (Figs. 6 and 7). Cladistic analysis was also conducted using a group average method. Except for Kai and Komoro, nine vineyards were very close to each other in the position map and eight vineyards formed a cluster in the cladogram (Fig. 6A). The four cultivars in the vineyards tested were widely scattered in the position map (Fig. 6B). On the other hand, Koshu and Pinot Noir at the shoot elongation stage, cultivated in Kofu, were close to each other in the position map and formed a cluster in the cladogram (Fig. 7A). Chardonnay and Cabernet Sauvignon at the shoot elongation stage, cultivated in Kofu, were close to each other but apart from Koshu and Pinot Noir, and formed a cluster in the cladogram. Interestingly, at véraison, the four cultivars were very close to each other in the position map (Fig. 7B).Figure 6Multidimensional scaling analysis of endophytic bacterial microbiota in grapevine shoot xylems among vineyards or cultivars. (A) Vineyards. (B) Cultivars. Left, position map. Right, cladogram. UR Urausu, MS Minamisanriku, KF Kofu, KA Kai, KN Katsunuma, KM Komoro, UE Ueda, SH Shobara, IZ Izumo, SA Saijo, OM Omishima, KO Koshu, CH Chardonnay, CS Cabernet Sauvignon, PN Pinot Noir.Full size imageFigure 7Multidimensional scaling analysis of endophytic bacterial microbiota in grapevine shoot xylems among cultivars grown in Kofu vineyard. (A) Shoot elongation stage. (B) Véraison. Left, position map. Right, cladogram. KO Koshu, CH Chardonnay, CS Cabernet Sauvignon, PN Pinot Noir.Full size imageNext, MDS and cladistic analysis of each cultivar in the vineyards were performed (Fig. 8). The distances among vineyards cultivating Koshu were small irrespective of the shoot growth stage (Fig. 8A). The distances among vineyards cultivating Pinot Noir were also small at the shoot elongation stage, and were further decreased at véraison (Fig. 8B). In contrast, the distances among vineyards cultivating Chardonnay and Cabernet Sauvignon were large at the shoot elongation stage (Fig. 8C,D). Although the distances among some vineyards (Urausu, Kofu, Kai, Izumo, and Shobara for Chardonnay, and Ueda, Kofu, Katsunuma, Komoro, and Shobara for Cabernet Sauvignon) decreased at véraison, they were large compared with Koshu and Pinot Noir.Figure 8Multidimensional scaling analysis of endophytic bacterial microbiota in grapevine shoot xylems among vineyards cultivating each cultivar. (A) Koshu. (B) Pinot Noir. (C) Chardonnay. (D) Cabernet Sauvignon. Left, position map. Right, cladogram. UR Urausu, MS Minamisanriku, KF Kofu, KA Kai, KN Katsunuma, KM Komoro, UE Ueda, SH Shobara, IZ Izumo, SA Saijo, OM Omishima.Full size image More

WoSIS and permafrost-affected soil profilesThe World Soil Information Service (WoSIS) collates and manages the largest database of explicit soil profile observations across the globe29. In this study, we used the quality-assessed and standardised snapshot of 2019 (ISRIC Data Hub). We further screened the snapshot, and excluded soil profiles with obvious errors (e.g., negative depth values of mineral soil, the value of the depth for the deeper layer is smaller than that of the upper layer). Finally, there is a total of 110,695 profiles with records of SOC content (SOCc, g C kg–1 soil) in the fine earth fraction < 2 mm. The soil layer depths are inconsistent between soil profiles. We harmonised SOCc to three standard depths (i.e., 0–0.3, 0.3–1 and 1–2 m) using mass-preserving splines61,62, which makes it possible to directly compare among soil profiles. We also calculated SOC stock (SOCs, kg C m–2) in each standard depth as:$${{{{{{rm{SOC}}}}}}}_{{{{{{rm{s}}}}}}}=frac{{{{{{{rm{SOC}}}}}}}_{{{{{{rm{c}}}}}}}}{100}cdot Dcdot {{{{{rm{BD}}}}}}cdot left(1-frac{G}{100}right),$$ (1) where D is the soil depth (i.e., 0.3, 0.7, or 1 m in this study), BD is the bulk density of the fine earth fraction 2 mm) of soil. Amongst the 110,695 soil profiles, unfortunately, only 18,590 profiles have measurements of both BD and G. To utilise and take advantage of all SOCc measurements, we used generalised boosted regression modelling (GBM) to perform imputation (i.e., filling missing data). As such, SOCs can be estimated. To do so, for BD and G in each standard soil depth, GBM was developed based on all measurements of that property (e.g., BD) in the 110,695 profiles with other 32 soil properties recorded in the WoSIS database. The detailed approach for missing data imputation has been described in ref. 41.Together with the WoSIS soil profiles, a total of 2,703 soil profiles with data of SOCs from permafrost-affected regions were obtained from ref. 30. The original data used in ref. 30 have been obtained, and we used the data of SOCs in the 0–0.3, 0.3–1, and 1–2 m soil layers in this study. These permafrost-affected profiles compensate for the scarce soil profiles in high latitudinal regions in the WoSIS database. Overall, the soil profiles cover 13 major biome groups although the profile numbers vary among biome types (Supplementary Fig. 1). The profiles also cover various climate conditions across the globe with mean annual temperature (MAT) ranging from –20.0 to 30.7 °C and mean annual precipitation (MAP) ranging from 0 to 6,674 mm.Environmental covariatesMAT and MAP for each soil profile were obtained from the WorldClim version 2 (ref. 63). The WorldClim version 2 calculates biologically meaningful variables using monthly temperature and precipitation during the period 1970–2000. We obtained global spatial layers of MAT and MAP at the resolution of 30 arcsecond (i.e., 0.0083° which is equivalent to ~1 km at the equator). Soil profiles in the same 0.0083° grid (i.e., ~1 km2) share the same MAT and MAP. Besides MAT and MAP, other climatic variables for each soil profile were also obtained from the WorldClim version 2. The WWF (World Wildlife Fund) map of terrestrial ecoregions of the world (https://www.worldwildlife.org/publications/terrestrial-ecoregions-of-the-world) was used to extract the biome type at each soil profile. The MODIS land cover map64 at the same resolution of NPP databases was used to identify that if the land is cultivated (i.e., land cover type of croplands and cropland/natural vegetation mosaic) at the location of each soil profile.Space-for-time substitution: grouping soil profilesWe used a hybrid approach of space-for-time substitution and meta-analysis to estimate the response of SOC to warming. Traditionally, space-for-time substitution involves determining regression relationships across gradients at one time31. The regression was then used to predict future status under conditions when one or more of the covariates has changed31. However, the approach was compromised when the effects of other driving variables such as soil type and landform were not minimised. Regarding SOC dynamics, they would show non-linear relationships19 with temperature modulated by a series of other environmental covariates (e.g., precipitation, vegetation type).Based on the idea of space-for-time approach31, first, we sorted all soil profiles by MAT at the soil-profile locations and designated them into MAT classes with an increment of 1 °C (Fig. 1). Then, we derived pairs of soil profiles, with each pair including a “ambient” and “warm” class (i.e., control vs treatment in meta-analysis language) distinguished by MAT (Fig. 1). The ambient class includes soil profiles with MAT ranging from i to i + 1 degree Celsius, where i is the lowest temperature in the class. If 1 °C warming is of interest, for example, the warm class will be identified as the class with MAT ranging from i + 1 to i + 2 degree Celsius (i.e., one degree higher than that of the ambient class; Fig. 1). To control the effects of precipitation, soil type and topography, soil profiles in both ambient and warm classes were further grouped; and each group must have the same following characteristics: (1) Landform. A global landform spatial layer was obtained from Global Landform classification - ESDAC - European Commission (europa.eu), and global terrestrial lands were divided into three general landform types: plains including lowlands, plateaus, and mountains including hills. (2) Soil type. The 12 USDA soil orders were used to distinguish soil types. A global spatial layer of soil orders was obtained from The Twelve Orders of Soil Taxonomy | NRCS Soils (usda.gov). We also independently tested the sensitivity of the results to different soil classification systems by including FAO and WRB soil groups (Soil classification | FAO SOILS PORTAL|Food and Agriculture Organization of the United Nations). (3) Mean annual precipitation (MAP). MAP cannot be exactly the same between the ambient and warm groups. In practice, we considered that soils meet this criterion if the absolute difference of MAP between ambient and warm soils is less than 50 mm. We also tested the sensitivity of the results to this absolute MAP difference using another value of 25 mm, and found that this difference has negligible effect (Supplementary Fig. 11). (4) Precipitation seasonality. Precipitation seasonality indicates the temporal distribution of precipitation. In this study, we focused on warming alone, and global warming would also have less effect on this seasonal distribution of precipitation. The seasonal distribution pattern of precipitation was classified into three categories: summer-dominated precipitation, winter-dominated precipitation and uniform precipitation. Precipitation concentration index (PCI) was calculated in R precintcon package to distinguish the three patterns65: $${{{{{rm{PCI}}}}}}=frac{mathop{sum }nolimits_{{{{{{rm{i}}}}}}=1}^{12}{p}_{{{{{{rm{i}}}}}}}^{2}}{{left(mathop{sum }nolimits_{{{{{{rm{i}}}}}}=1}^{12}{p}_{{{{{{rm{i}}}}}}}right)}^{2}}cdot 100,$$ (2) where pi is the precipitation in month i in a particular year. In this study, we used the monthly precipitation from 1970 to 2000 obtained from WorldClim version 2 (ref. 63) to calculate the average (overline{{{{{{rm{PCI}}}}}}}) at the location of each profile. If (overline{{{{{{rm{PCI}}}}}}})  8.3 and total precipitation from April to September (from October to March in the Southern Hemisphere) is larger than that from October to March (from April to September in the Southern Hemisphere), precipitation mainly occurs in summer (i.e., summer precipitation); otherwise, it is winter precipitation.By applying these selection criteria to all soil profiles, we obtained pairs (i.e., an “ambient” group vs a “warm” group) of soil profiles mainly distinguished by MAT (i.e., warming). Amongst pairs, they would be different in landform, soil type, MAP and precipitation seasonality, which enables us to address their effects on the response of SOC to warming. We are interested in five warming levels including 1, 2, 3, 4, and 5 °C.Meta-analysis: estimation of the response of SOC to warmingMeta-analysis techniques were used to estimate the percentage response of SOC to warming by comparing SOC content and stock in groups in the warm group to that in the ambient group. The log response ratio of soil C (lnRR) to warming for each pair (i.e., an ambient group vs a warm group) of soil profiles was calculated as:$${{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}={{{{{rm{ln}}}}}}left(frac{bar{{{{{{{rm{SOC}}}}}}}^{*}}}{overline{{{{{{rm{SOC}}}}}}}}right),$$ (3) where (overline{{{{{{rm{SOC}}}}}}}) and (bar{{{{{{{rm{SOC}}}}}}}^{*}}) are the mean SOC (either content or stock) in groups from ambient and warm class, respectively. In order to provide a robust estimate of global mean response ratio, the individual lnRR values were weighted by the inverse of the sum of within- (v) and between-group (τ2) variances. As such, the global mean response ratio ((overline{{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}})) could be estimated as:$$overline{{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}}=frac{{sum }_{{{{{{rm{i}}}}}}}left({{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}}_{{{{{{rm{i}}}}}}}times {w}_{{{{{{rm{i}}}}}}}right)}{{sum }_{{{{{{rm{i}}}}}}}{w}_{{{{{{rm{i}}}}}}}},$$ (4) where ({w}_{{{{{{rm{i}}}}}}}=frac{1}{{v}_{{{{{{rm{i}}}}}}}+{tau }^{2}}) is the weight for the ith lnRR. In addition, we estimated and compared the mean response ratios under different soil orders, landforms, and precipitation concentration patterns. These mean response rates were calculated in weighted, mixed-effects models using the rma.mv function in R package metafor. To assist interpretation, the results of (overline{{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}}) were back-transformed and reported as percentage change under warming, i.e., (({{{{{{rm{e}}}}}}}^{{{{{{rm{RR}}}}}}}-1)times)100. These back-transformed values were also used for subsequent data analyses.An implicit assumption underlying the space-for-time substitution approach is that important events or processes which substantially change the succession direction of studied system (e.g., volcano disruption in one class but not in another class, cultivation in one class but not in another class) are independent of space and time (which includes the past and future)66. We conducted two sensitivity assessment to test this assumption. First, we repeated all above assessment by excluding soil profiles from croplands since preferential choice of land clearing for cultivation should be common. Second, we repeated all assessment by including only groups having at least 20 soil profiles. This allows the assessed pairs to cover a higher diversity of land history and future land cover/use, diluting the effect of a typical event at a specific soil profile on the estimates.Comparison with SOC turnover modelsWe compared our estimation with predictions by SOC models. A simple one-pool SOC model can be written as:$$frac{{{{{{rm{d}}}}}}C}{{{{{{rm{d}}}}}}t}=I-kcdot C,$$ (5) where I is the amount of carbon input, k is the decay rate of SOC, and C is the stock of SOC. At steady state, (C=I/k). A Q10 function can be applied to estimate k under warming (kw):$${k}_{{{{{{rm{w}}}}}}}=kcdot {{exp }}left(0.1cdot triangle Tcdot {{log }}left({Q}_{10}right)right),$$ (6) where (triangle T) is the warming level. Thus, when soil reaches a new steady state under warming, SOC stock (Cw) can be estimated as:$${C}_{{{{{{rm{w}}}}}}}=frac{{I}_{{{{{{rm{w}}}}}}}}{kcdot {{exp }}left(0.1cdot triangle Tcdot {{log }}left({Q}_{10}right)right)},$$ (7) where Iw is the carbon input amount under warming condition. Finally, the response of SOC to warming (R) can be calculated as:$$R=frac{{C}_{{{{{{rm{w}}}}}}}-C}{C}=frac{{I}_{{{{{{rm{w}}}}}}}}{I}cdot {{exp }}left(-0.1cdot triangle Tcdot {{log }}left({Q}_{10}right)right)-1.$$ (8) Using Eq. (8), we calculated R under a series of ensembles of (frac{{I}_{{{{{{rm{w}}}}}}}}{I}), (triangle T), and ({Q}_{10}), and compared R with that estimated using our space-for-time substitution approach.Comparison with field warming experimentsA number of meta-analyses based on data from field warming experiments had been performed to assess the response of SOC to warming7,26,46,47,48,49,50, which enable us to conduct comparisons with the estimates using our hybrid approach combining space-for-time substitution and meta-analysis techniques. A total of five meta-analysis papers have been found by searching the Web of Science. We retrieved the response ratios from the identified papers, and compared them to our estimations. Here, it should be noted that most field warming experiments focused on SOC changes (stock or content) in the top 0.2 m soil layer. We compared them with our estimation of the response of SOC stock in the top 0.3 m soil.Besides the published results of meta-analysis, we also conducted an independent meta-analysis using data from field warming experiments. The meta-analysis dataset was mainly from published papers on meta-analysis from 2013 to 2020 (see Supplementary Data 1). It should be noted that the field warming experiments manipulate temperature using different approaches such as open/closed-top chamber, infrared radiators and heating cables. For the comparison, we did not explicitly distinguish these approaches. The experimental duration ranged from 0.42 to 25 years with a mean of 4.7 years, and the warming magnitude ranged from 0.1 to 7°C with a mean of 1.92 °C. To ease comparison, field warming levels were classified into 0–1, 1–2, 2–3, 3–4, 4–5, and >5 °C. The same meta-analysis to that assessing soil profile data was used to predict the response ratio of SOC to the above six warming levels. In addition, we divided the data into four ecosystems (i.e., tundra, forest, shrublands and grasslands) and estimated the response ratio in each ecosystem. These estimates based on field warming experiments were compared with those estimated using our space-for-time approach.Variable importance and global mappingWe included 15 environmental predictors to derive a meta-forest model, a machine learning-based random forest model adapted for meta-analysis, to map the response of SOC stock/content to warming across the globe at the resolution of 0.0083°. The 15 environmental predictors reflect generally four broad groups of environmental conditions: baseline SOC conditions represented by current standing SOC stock or content, soil order and soil depth; current baseline climatic conditions represented by MAT, MAP, aridity index, precipitation seasonality represented by PCI, the fraction of precipitation in summer, the difference of temperature between ambient and warm groups, the difference of precipitation between ambient and warm groups; topography represented by elevation and landform; and vegetation represented by NPP and biome type.The metaforest function in the metafor package was used to derive the model. To fit the model, a fivefold cross-validation was conducted. That is, 80% of the derived response ratios was used to train the model, and the remaining 20% to validate the model. The best model hypeparameters were targeted by running the model under a series of parameter combinations, and the model performance was assessed by the rooted mean squared error (RMSE) and determination coefficient (R2). The meta-forest model allows the estimation of the relative influence of each individual variable in predicting the response, i.e. the relative contribution of variables in the model. The relative influence is calculated based on the times a variable selected for splitting when growing a tree, weighted by squared model improvement due to that splitting, and then averaged over all fitted trees which are determined by the algorithm when adding more trees cannot reduce prediction residuals. As such, the larger the relative influence of a variable, the stronger the effect of the variable on the response variable.Combining with spatial layers of predictors, the meta-forest model for SOC stock was used to predict the response of SOC to warming across the globe at the resolution of 1 km (most data layers are already at the 1 km resolution as abovementioned, for those layers that are not at the target resolution, they were resampled to the 1 km resolution). In the meta-forest model, current standing SOC stock is the most important predictor (Fig. 4). We use three global maps of SOC stocks including WISE51 (WISE Soil Property Databases | ISRIC), HWSD52 (Harmonized World Soil Database (HWSD v 1.21) – HWSD – IIASA) and SoilGrids53 (SoilGrids250m 2.0) to obtain current standing SOC stocks. These three global maps represent the major mapping products of SOC stock at the global level, and had been widely used for large scale modelling. The derived meta-forest model was applied across the globe to estimate the response ratio of SOC stock in each 1 km pixel. To do so, the same procedure to group the observed soil profiles (Fig. 1) was applied to group global land pixels (section Space-for-time substitution: grouping soil profiles). The only difference is that global mapping uses all pixels instead of the 113,013 soil profiles. In each 1 km pixel, prediction uncertainty was also quantified using estimates of randomly drawn 500 trees of the fitted meta-forest model to calcuate standard deviation of the predictions. More

The European Union’s new (2023–2027) Common Agricultural Policy (CAP) aims to reverse current environmental degradation and biodiversity declines in European farmland1 through the achievement of three green objectives: contribute to climate change mitigation, support efficient natural resource management, and reverse biodiversity loss2,3. Following the outbreak of war in Ukraine, the European Commission proposed a series of short and medium-term relaxations to CAP’s environmental commitments to offset expected shortages in grain imports and enhance food security4.Here, we argue that policy changes to allow cultivation of fallow land will disproportionately impact biodiversity and support further intensification of livestock production. Thus, ultimately, these changes in policy may sacrifice long term biodiversity and agricultural sustainability in Europe, in favour of modest increases in current agricultural production and alleged improvements of food security.A catalyst for reversing green policiesRussia and Ukraine are world-leading producers and exporters of cereal and fodder production (notably, oleo-proteaginous crops)5. The Ukraine war and international sanctions on Russia are threatening the import of these products to the EU. Ukrainian winter cereal, maize and sunflower production is expected to decrease by 20–30%, at least during the 2022–2023 season, and similar reductions in Russian exports are also expected5. Therefore agro-industry lobbies and farmers’ organisations in Brussels, some political parties in the European parliament and some countries’ administrations perceive a need to increase agricultural production6 and, as a means to offset expected shortages, are pressing to relax or remove CAP’s environmental commitments. Mechanisms supporting these commitments include enhanced conditionality (compulsory for all farmers receiving subsidies), voluntary measures of Rural Development Programmes (i.e. agri-environment-climate-measures) and Greening measures (crop diversification, maintenance of permanent grasslands and promotion of Ecological Focus Areas). A call made to mobilise all relevant international groups during the informal meeting held on 2 March 2022 by Member States’ agriculture and food ministers, with the exception of Denmark, Germany and Italy, may reflect such pressure6. Indeed, the European Commission has finally proposed a series of “short- and medium-term actions to enhance global food security and to support farmers and consumers in the EU”4. In regard to land-use, actions refer to the cultivation of fallows, which are protected by green payments for keeping land in good agricultural and environmental conditions and, adequately managed (both long-term and annual), support high levels of biodiversity and ecosystem services7 (Fig. 1). More precisely, the European Commission proposes that “To enlarge the EU’s production capacity, the Commission has today adopted an implementing act to exceptionally and temporarily allow Member States to derogate from certain greening obligations. In particular, they may allow for production of any crops for food and feed on fallow land that is part of Ecological Focus Areas in 2022, while maintaining the full level of the greening payment”4. This measure was recently extended for 2023.Fig. 1: Arable field left fallow and allowed to develop a grassy vegetation cover.Under non-intensive management, fallow areas become a genuine semi-natural habitat, key for the conservation of farmland biodiversity. Credit: Jordi Bas, taken in the cereal steppes of the Lleida plain (Catalonia, Spain).Full size imageConsidering food sovereigntyHowever, the FAO does not draw the same conclusions about the possible world impacts of the conflict and recommends finding alternative suppliers, instead suggesting using existing food stocks, diversifying domestic crops and reducing fertiliser dependence and food waste as mechanisms to help guarantee Europe’s food supplies and sovereignty5. Even the European Commission, while acknowledging the vulnerability of European farmers to animal feed import shortages and increased costs, clearly stated that food supply is not at risk in the EU4. Indeed, EU-based production supplies 79% of the feed proteins consumed in European livestock farming, 90% of feed cereals and 93% of other products such as Dried Distillers’ Grains and Solubles or beet pulp8. In 2020, the EU was completely self-sufficient with respect to dairy products, pork, beef, veal, poultry, and cereals. It remained the largest global exporter of agri-food products, in spite of the COVID-19 pandemic8.Counterproductive policiesAny increase in production from cultivating fallow land will therefore likely be used to feed intensively reared livestock and sustain cattle feed exports. Supporting the increasing trend of feed exports and industrial intensive livestock farming does not align with the EU’s Green Deal due to the negative impacts on air, soil and water quality8,9,10. In addition, cultivating fallow land to support intensive livestock-based agriculture will undermine the EU’s Farm-to-Fork strategy and CAP’s ‘Food and Health’ objective of reducing meat consumption to favour a more sustainable and healthier diet among European consumers2,11. Encouraging the growth of intensive livestock farming through enabling cultivation of fallow lands will increase environmental damage, biodiversity loss and public health risks. Thus, the recent relaxations of the new CAP compromise several of its fundamental objectives, along with those of other elements of the Green Deal, such as the EU’s Nature Restoration Law2,9,12.The duration of the war in Ukraine and its effects on provision of raw materials to Europe is hard to foresee. We acknowledge the uncertainties and input costs faced by farmers but calls for further agricultural intensification may be largely unjustified at this stage. Specifically, cultivating semi-natural habitats like long-term or unploughed annual fallows will have serious environmental costs, including an increase in pesticide and fertiliser application, since fallows often occupy less productive land13. Even a moderate increase in food production at the expense of the semi-natural habitats remaining in farmland landscapes (field margins, grasslands, and fallow land), which support most of Europe’s farmland biodiversity and its associated ecosystem services14, will seriously damage farmland biodiversity and sustainability in European agricultural landscapes3,15. For example, a comprehensive study carried out on 169 farms across 10 European countries showed that semi-natural habitats, including fallows, occupied 23% of the land but hosted 49% of vascular plant, earthworm, spider, and wild bee species; a 10% decrease of these habitats if reclaimed for food production would cause exponential decreases in biodiversity, but only moderate linear increases in production15. Furthermore, the loss of semi-natural habitats in arable systems, fallows among them, would negatively affect arthropod functional diversity and the ecosystem services it supports, which may affect agriculture production14.Sustainable alternativesThere are alternatives to cultivating semi-natural habitats that may (and need to) be assessed to achieve a more strategic European agricultural policy able to meet food demands while maintaining the sustainability principles and improvements of the food-production chain sought by the Farm-to-Fork strategy. Proposals include agro-ecological approaches to increase production through the enhancement of ecosystem services such as pollination and biological control16,17,18, adjusting the amount of cultivated surface in relation to landscape structure and composition19, or relocating crops that are more in demand to areas where production is optimal without increasing the total cultivated area20.After decades of costly implementation and reforms of agricultural and conservation policies1, the EU is at risk of engaging in a hasty and misguided strategy on food production jeopardising the green transition13. As an alternative to such a ‘business as usual’ reaction, the EU has now the opportunity to consolidate the mentioned environmental and social objectives of the new CAP2,3. A more sustainable agriculture, resilient to food supply crises (present and future), should be based on ecological functionality of farmland, which ultimately depends on the conservation of its biodiversity16, along with measures to counter climate change. Responses to this and other challenges on the new CAP should be assessed with a long-term perspective and based on robust scientific evidence before undermining its environmental ambitions3,13. More

Study areaSoutheastern Chongqing, China (107° 14′–109° 19′ E, 28° 9′–30° 32′ N), has an area of about 19,800 km2 (Fig. 1). The study area has a subtropical monsoon climate. And the area has four distinct seasons, with an annual average temperature of 16.2 °C and abundant rainfall, with an average annual rainfall of 1209 mm. This region is located in the central part of the Wuling mountains, which is characterized by medium and low mountainous landforms, with an average altitude of greater than 1000 m. The water system (the Wujiang River system) in the study area is well developed, with a large drainage area and rich groundwater resources. The soil is dominated by yellow soil and limestone soil, and the sensitivity to soil erosion is high. The district exhibits the typical ecological fragility of karst areas, with barren soil, fragmented surfaces, a single community, and a low ecological carrying capacity. The area includes six counties: Qianjiang district, Shizhu Tujia Autonomous county, Xiushan Tujia and Miao Autonomous county, Youyang Tujia and Miao Autonomous county, Wulong district, and Pengshui Miao and Tujia Autonomous county. The coverage rates of the carbonatite layers in these counties are 42.11, 67.77, 25.70, 34.80, 59.70 and 88.46%, respectively38, and the average coverage of the carbonatite layers is 53.09%, making this a representative area of karst rocky desertification.Data and image pre-processingIn the study, the remote sensing data were obtained from the United States Geological Survey (USGS, https://earthexplorer.usgs.gov/), including landsat-5 thematic mapper (TM) images acquired in 2001, 2006 and 2011 and Landsat-8 operational land imager (OLI) images obtained in 2016 and 2021 (Table 1). The spatial resolution is 30 m. In order to ensure the comparability of spectral characteristics, the data collection was conducted from May to September when the vegetation grew better. In order to meet the usage requirements, the cloud cover of each image used is below 10%. For the images with poor quality, the adjacent years were selected for replacement. The difference in ecological quality between adjacent years in the same region was not particularly large. In order to represent the actual situation of the ecological environment quality in the target year as much as possible, we tried to minimize the replaced part in each target year. A total of 20 images were collected in this study. The images downloaded were all L1T products, which had undergone systematic radiometric correction and geometric correction, so precise geometric correction was no longer performed. Before the subsequent processing, all 20 images were preprocessed by radiometric calibration, atmospheric correction, image mosaicking and cropping. Then these images were calculated to obtain NDVI, WET, NDBSI, LST and RI. And based on the preprocessed Landsat images, support vector machine classification was performed to obtain the land use (LU) status.Table 1 Information of images used in this study.Full size tableThe topographical data included the elevation (EV) and slope (SP) data. Among them, the elevation data was provided by the official website of the United States Geological Survey (USGS, https://earthexplorer.usgs.gov/). And the slope data was calculated from the elevation data. The meteorological data, including the monthly average temperature (MT), monthly mean precipitation (PR), monthly even relative humidity (RH), and monthly total sunshine hours (SH) from May to September of the target year, were got from the China Meteorological Data Network (http://data.cma.cn/). In addition, socioeconomic data, including the population density (PD) and gross domestic product (GDP), were obtained from the statistical yearbooks of each district and county in the study area. The nighttime light (NTL) data were obtained from the National Oceanic and Atmospheric Administration (NOAA, https://www.noaa.gov/). The above data and LU were used as the influencing factors of ecological quality to analyze the reasons for the change of local ecological environment quality. The statistical data and monitoring data of each evaluation index used to construct the EI come from the statistical yearbooks, water resources bulletin and soil and water conservation bulletin of each district and county.MethodologyStudy frameworkA framework was developed for evaluating the ecological quality in southeastern Chongqing from 2001 to 2021 in the study. And the framework included three parts: data preparation, construction of the MRSEI, and the analysis of the ecological status in the region. Figure 2 presents the detailed information about the framework. The operations of band calculation, normalization and PCA were all carried out using the ENVI 5.3 software (https://www.harrisgeospatial.com).Figure 2The study framework.Full size imageIndicators used in MRSEIThe greenness, humidity, heat, dryness, and degree of rocky desertification were used to construct the MRSEI. The NDVI39 was chosen to characterize the greenness. The humidity component acquired from the tasseled cap transformation (WET)40 was selected to represent the humidity. The LST41 was used to represent the heat, the normalized difference build-up soil index (NDBSI)42 was used to characterize the dryness. The RI was applied to characterize the degree of rocky desertification.The NDVI is an important indicator for monitoring the physical and chemical properties of vegetation, and it can be employed to calculate the vegetation coverage, leaf area index, and so on19. In addition, it eliminates some radiation errors and has a stronger response to surface vegetation. It has been widely used in vegetation remote sensing monitoring. The equation for calculating the NDVI is as follows39:$$ {text{NDVI}} = {{(uprho }}_{{{text{NIR}}}} – {uprho }_{{{text{Red}}}} {)}/{{(uprho }}_{{{text{NIR}}}} {{ + uprho }}_{{{text{Red}}}} ), $$
(1)
where ({uprho }_{{{text{NIR}}}}) is the reflectance of the near-infrared band and ({uprho }_{{{text{Red}}}}) refers to the reflectance of the red band corresponding to each image.The WET can effectively reflect the humidity conditions of the surface vegetation, water, and soil, and can reveal the changes in the ecological environment, such as soil degradation. Therefore, it is commonly used in ecological environment monitoring43. The WET can be expressed as40,43:$$ {text{WET}}_{{{text{TM}}}} { = 0}{{.3102uprho }}_{{{text{Red}}}} { + 0}{{.2021uprho }}_{{{text{Green}}}} { + 0}{{.0315uprho }}_{{{text{Blue}}}} { + 0}{{.1594uprho }}_{{{text{NIR}}}} – {0}{{.6806uprho }}_{{{text{SWIR1}}}} – {0}{{.6109uprho }}_{{{text{SWIR2}}}} , $$
(2)
$$ {text{WET}}_{{{text{OLI}}}} { = 0}{{.3283uprho }}_{{{text{Red}}}} { + 0}{{.1972uprho }}_{{{text{Green}}}} { + 0}{{.1511uprho }}_{{{text{Blue}}}} { + 0}{{.3407uprho }}_{{{text{NIR}}}} – {0}{{.7117uprho }}_{{{text{SWIR1}}}} – {0}{{.4559uprho }}_{{{text{SWIR2}}}} , $$
(3)
where ({uprho }_{{text{i}}} ,) is the reflectance of band i.The NDBSI is expressed as the average of two indicators, the bare soil index (SI)44 and the index-based built-up index (IBI)45. It can be applied to characterize the dryness. The calculation formulas are44,45:$$ {text{IBI }} = {text{ }}left[ {2uprho _{{{text{SWIR1}}}} /left( {uprho _{{{text{SWIR1}}}} + {text{ }}uprho _{{{text{NIR}}}} } right) – uprho _{{{text{NIR}}}} /(uprho _{{{text{NIR}}}} + {text{ }}uprho _{{{text{Red}}}} } right) – uprho _{{{text{Green}}}} /(uprho _{{{text{Green}}}} + {text{ }}uprho _{{{text{SWIR1}}}} )]/[2uprho _{{{text{SWIR1}}}} /left( {uprho _{{{text{SWIR1}}}} + {text{ }}uprho _{{{text{NIR}}}} } right) + {text{ }}uprho _{{{text{NIR}}}} /(uprho _{{{text{NIR}}}} + {text{ }}uprho _{{{text{Red}}}} ) + {text{ }}uprho _{{{text{Green}}}} /(uprho _{{{text{Green}}}} + {text{ }}uprho _{{{text{SWIR1}}}} )], $$
(4)
$$ {text{SI = }}left[ {{uprho }_{{{text{SWIR1}}}} {{ + uprho }}_{{{text{red}}}} – left( {{uprho }_{{{text{Blue}}}} {{ + uprho }}_{{{text{NIR}}}} } right)} right]/left[ {{uprho }_{{{text{SWIR1}}}} {{ + uprho }}_{{{text{red}}}} { + }left( {{uprho }_{{{text{Blue}}}} {{ + uprho }}_{{{text{NIR}}}} } right)} right], $$
(5)
$$ {text{NDBSI = (IBI + SI)/2,}} $$
(6)
where ({uprho }_{{text{i}}} ,) is the reflectance of band i.The LST is closely related to natural processes and human phenomena such as crop yield, vegetation growth and distribution, surface water cycle, etc. It can well reflect the state of the surface ecological environment. The atmospheric correction method is used to invert the LST here46,47, it can be expressed as:$$ {text{L = gain}} times {text{DN + bias,}} $$
(7)
$$ {text{T = K}}_{{2}} /{text{ln}}left( {frac{{{text{K}}_{{1}} }}{{text{L}}}{ + 1}} right){,} $$
(8)
$$ {text{LST = T}}/left[ {{1 + }left( {frac{{{lambda T}}}{{upalpha }}} right){{lnvarepsilon }}} right]{,} $$
(9)
where L is the radiation value in the thermal infrared band, DN is the gray value, gain and bias is the gain value and offset value of the L-band, which was got from the image header file. And T is the temperature value at the sensor; K1 and K2 are calibration parameters respectively (for TM, K1 = 607.76 W/(m2 sr μm), K2 = 1260.56 K; for TIRS, K1 = 774.89 W/(m2 sr μm), K2 = 1321.08 K); λ is the central wavelength of thermal infrared band; α = 1.438 × 10−2 m K. ε is the surface emissivity and the value is estimated by the vegetation index mixture model48,49. It is calculated as follows:$$ {text{VFC = }}frac{{{text{NDVI}} – {text{NDVI}}_{{{text{Soil}}}} }}{{{text{NDVI}}_{{{text{Veg}}}} – {text{NDVI}}_{{{text{Soil}}}} }}, $$
(10)
$$ {text{d}}_{{upvarepsilon }} { = }left( {{1} – {upvarepsilon }_{{text{s}}} } right){{ times (1}} – {text{VFC) }}times text{F} times upvarepsilon _{{text{v}}} , $$
(11)
$$ {{upvarepsilon = upvarepsilon }}_{{text{v}}} times {text{ VFC}} + varepsilon _{{text{s}}} {{ times }}left( {{1} – {text{FVC}}} right){text{ + d}}_{{upvarepsilon }} , $$
(12)
where VFC is the vegetation fractional cover, ({text{NDVI}}_{{{text{Veg}}}}) is the NDVI of the pixel covered by full vegetation and the pixels with NDVI  > 0.72 are regarded as pure vegetation pixels; ({text{NDVI}}_{{{text{Soil}}}}) is the NDVI of the bare pixel and the pixels with NDVI  More