Ecology

Fuhrman JA, Campbell L. Marine ecology – microbial microdiversity. Nature. 1998;393:410–1.CAS 

Google Scholar 
Thompson JR, Pacocha S, Pharino C, Klepac-Ceraj V, Hunt DE, Benoit J, et al. Genotypic diversity within a natural coastal bacterioplankton population. Science. 2005;307:1311–3.CAS 
PubMed 

Google Scholar 
Oh S, Buddenborg S, Yoder-Himes DR, Tiedje JM, Konstantinidis KT. Genomic diversity of Escherichia isolates from diverse habitats. PLoS ONE. 2012;7:e47005.CAS 
PubMed 
PubMed Central 

Google Scholar 
Fuhrman JA, Steele JA, Hewson I, Schwalbach MS, Brown MV, Green JL, et al. A latitudinal diversity gradient in planktonic marine bacteria. Proc Natl Acad Sci USA. 2008;105:7774–8.CAS 
PubMed 
PubMed Central 

Google Scholar 
Sunagawa S, Coelho LP, Chaffron S, Kultima JR, Labadie K, Salazar G, et al. Structure and function of the global ocean microbiome. Science. 2015;348:1261359.PubMed 

Google Scholar 
Raes EJ, Bodrossy L, van de Kamp J, Bissett A, Ostrowski M, Brown MV, et al. Oceanographic boundaries constrain microbial diversity gradients in the South Pacific Ocean. Proc Natl Acad Sci USA. 2018;115:EB266–EB75.
Google Scholar 
Brown MV, Fuhrman JA. Marine bacterial microdiversity as revealed by internal transcribed spacer analysis. Aquat Microb Ecol. 2005;41:15–23.
Google Scholar 
Johnson ZI, Zinser ER, Coe A, McNulty NP, Woodward EMS, Chisholm SW. Niche partitioning among Prochlorococcus ecotypes along ocean-scale environmental gradients. Science. 2006;311:1737–40.CAS 
PubMed 

Google Scholar 
Larkin AA, Blinebry SK, Howes C, Chandler J, Zinser ER, Johnson ZI. Niche partitioning and biogeography of high light adapted Prochlorococcus across taxonomic ranks in the North Pacific. ISME J. 2016;10:1555–67.PubMed 
PubMed Central 

Google Scholar 
Zinser ER, Johnson ZI, Coe A, Karaca E, Veneziano D, Chisholm SW. Influence of light and temperature on Prochlorococcus ecotype distributions in the Atlantic Ocean. Limnol Oceanogr. 2007;52:2205–20.
Google Scholar 
Carlson M, Ribalet F, Maidanik I, Durham BP, Hulata Y, Ferrón S, et al. Viruses affect picocyanobacterial abundance and biogeography in the North Pacific Ocean. Nat Microbiol. 2022;7:570–80.CAS 
PubMed 
PubMed Central 

Google Scholar 
Follett CL, Dutkiewicz S, Ribalet F, Zakem E, Caron D, Armbrust EV, et al. Trophic interactions with heterotrophic bacteria limit the range of Prochlorococcus. Proc Natl Acad Sci USA. 2022;119:e2110993118.CAS 
PubMed 
PubMed Central 

Google Scholar 
Moore LR, Rocap G, Chisholm SW. Physiology and molecular phylogeny of coexisting Prochlorococcus ecotypes. Nature. 1998;393:464–7.CAS 
PubMed 

Google Scholar 
Zwirglmaier K, Jardillier L, Ostrowski M, Mazard S, Garczarek L, Vaulot D, et al. Global phylogeography of marine Synechococcus and Prochlorococcus reveals a distinct partitioning of lineages among oceanic biomes. Environ Microbiol. 2008;10:147–61.PubMed 

Google Scholar 
Carlson CA, Morris R, Parsons R, Treusch AH, Giovannoni SJ, Vergin K. Seasonal dynamics of SAR11 populations in the euphotic and mesopelagic zones of the northwestern Sargasso Sea. ISME J. 2009;3:283–95.CAS 
PubMed 

Google Scholar 
Ribalet F, Swalwell J, Clayton S, Jimenez V, Sudek S, Lin YJ, et al. Light-driven synchrony of Prochlorococcus growth and mortality in the subtropical Pacific gyre. Proc Natl Acad Sci USA. 2015;112:8008–12.CAS 
PubMed 
PubMed Central 

Google Scholar 
Pernthaler A, Pernthaler J, Schattenhofer M, Amann R. Identification of DNA-synthesizing bacterial cells in coastal North Sea plankton. Appl Environ Microbiol. 2002;68:5728–36.CAS 
PubMed 
PubMed Central 

Google Scholar 
Del Giorgio PA, Gasol JM. Physiological structure and single-cell activity in marine bacterioplankton. In: Kirchman DL, editor. Microbial ecology of the oceans. 2nd ed. New Jersey:John Wiley & Sons, Inc.; 2008. p. 243–298.Lin Y, Gazsi K, Lance VP, Larkin A, Chandler J, Zinser ER, et al. In situ activity of a dominant Prochlorococcus ecotype (eHL-II) from rRNA content and cell size. Environ Microbiol. 2013;15:2736–47.CAS 
PubMed 

Google Scholar 
Kirchman DL. Growth rates of microbes in the oceans. Annu Rev Mar Sci. 2016;8:285–309.
Google Scholar 
Landry MR, Selph KE, Yang EJ. Decoupled phytoplankton growth and microzooplankton grazing in the deep euphotic zone of the eastern equatorial Pacific. Mar Ecol Prog Ser. 2011;421:13–24.
Google Scholar 
Selph KE, Landry MR, Taylor AG, Yang E-J, Measures CI, Yang J, et al. Spatially-resolved taxon-specific phytoplankton production and grazing dynamics in relation to iron distributions in the Equatorial Pacific between 110 and 140 degrees W. Deep Sea Res Part II Top Stud Oceanogr. 2011;58:358–77.CAS 

Google Scholar 
Hunt DE, Lin Y, Church MJ, Karl DM, Tringe SG, Izzo LK, et al. Relationship between abundance and specific activity of bacterioplankton in open ocean surface waters. Appl Environ Microbiol. 2013;79:177–84.CAS 
PubMed 
PubMed Central 

Google Scholar 
Sintes E, Herndl GJ. Quantifying substrate uptake by individual cells of marine bacterioplankton by catalyzed reporter deposition fluorescence in situ hybridization combined with microautoradiography. Appl Environ Microbiol. 2006;72:7022–8.CAS 
PubMed 
PubMed Central 

Google Scholar 
Hall E, Maixner F, Franklin O, Daims H, Richter A, Battin T. Linking microbial and ecosystem ecology using ecological stoichiometry: a synthesis of conceptual and empirical approaches. Ecosystems. 2011;14:261–73.
Google Scholar 
Musat N, Foster R, Vagner T, Adam B, Kuypers MM. Detecting metabolic activities in single cells, with emphasis on nanoSIMS. FEMS Microbiol Rev. 2012;36:486–511.CAS 
PubMed 

Google Scholar 
Hatzenpichler R, Scheller S, Tavormina PL, Babin BM, Tirrell DA, Orphan VJ. In situ visualization of newly synthesized proteins in environmental microbes using amino acid tagging and click chemistry. Environ Microbiol. 2014;16:2568–90.CAS 
PubMed 
PubMed Central 

Google Scholar 
Sebastián M, Gasol JM. Visualization is crucial for understanding microbial processes in the ocean. Philos Trans R Soc Lond B. 2019;374:20190083.
Google Scholar 
Korem T, Zeevi D, Suez J, Weinberger A, Avnit-Sagi T, Pompan-Lotan M, et al. Growth dynamics of gut microbiota in health and disease inferred from single metagenomic samples. Science. 2015;349:1101–6.CAS 
PubMed 
PubMed Central 

Google Scholar 
Brown CT, Olm MR, Thomas BC, Banfield JF. Measurement of bacterial replication rates in microbial communities. Nat Biotech. 2016;34:1256–63.CAS 

Google Scholar 
Gao Y, Li H. Quantifying and comparing bacterial growth dynamics in multiple metagenomic samples. Nat Methods. 2018;15:1041–4.CAS 
PubMed 
PubMed Central 

Google Scholar 
Emiola A, Zhou W, Oh J. Metagenomic growth rate inferences of strains in situ. Sci Adv. 2020;6:eaaz2299.CAS 
PubMed 
PubMed Central 

Google Scholar 
Long ANM, Hou SW, Ignacio-Espinoza JC, Fuhrman JA. Benchmarking microbial growth rate predictions from metagenomes. ISME J. 2020;15:183–195.PubMed 
PubMed Central 

Google Scholar 
Carroll J, Van Oostende N, Ward BB. Evaluation of genomic sequence-based growth rate methods for synchronized Synechococcus cultures. Appl Environ Microbiol. 2021;01743-21.Vaulot D, Marie D. Diel variability of photosynthetic picoplankton in the equatorial Pacific. J Geophys Res Oceans. 1999;104:3297–310.CAS 

Google Scholar 
Hunter-Cevera KR, Neubert MG, Olson RJ, Shalapyonok A, Solow AR, Sosik HM. Seasons of Syn. Limnol Oceanogr. 2020;65:1085–102.PubMed 

Google Scholar 
Baer SE, Rauschenberg S, Garcia CA, Garcia NS, Martiny AC, Twining BS, et al. Carbon and nitrogen productivity during spring in the oligotrophic Indian Ocean along the GO-SHIP IO9N transect. Deep Sea Res Part II Top Stud Oceanogr. 2019;161:81–91.CAS 

Google Scholar 
Larkin AA, Garcia CA, Ingoglia KA, Garcia NS, Baer SE, Twining BS, et al. Subtle biogeochemical regimes in the Indian Ocean revealed by spatial and diel frequency of Prochlorococcus haplotypes. Limnol Oceanogr. 2020;65:S220–S32.CAS 

Google Scholar 
Larkin AA, Garcia CA, Garcia N, Brock ML, Lee JA, Ustick LJ, et al. High spatial resolution global ocean metagenomes from Bio-GO-SHIP repeat hydrography transects. Sci Data. 2021;8:1–6.
Google Scholar 
Baym M, Kryazhimskiy S, Lieberman TD, Chung H, Desai MM, Kishony R. Inexpensive multiplexed library preparation for megabase-sized genomes. PLoS ONE. 2015;10:e0128036.PubMed 
PubMed Central 

Google Scholar 
Wandro S, Oliver A, Gallagher T, Weihe C, England W, Martiny JBH, et al. Predictable molecular adaptation of coevolving Enterococcus faecium and lytic phage EfV12-phi1. Front Microbiol. 2019;9:3192.PubMed 
PubMed Central 

Google Scholar 
Oliver A, LaMere B, Weihe C, Wandro S, Lindsay KL, Wadhwa PD, et al. Cervicovaginal microbiome composition is associated with metabolic profiles in healthy pregnancy. mBio. 2020;11:e01851–20.CAS 
PubMed 
PubMed Central 

Google Scholar 
Eren AM, Esen OC, Quince C, Vineis JH, Morrison HG, Sogin ML, et al. Anvi’o: an advanced analysis and visualization platform for ‘omics data. PeerJ. 2015;3:e1319.PubMed 
PubMed Central 

Google Scholar 
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol. 1990;215:403–10.CAS 
PubMed 

Google Scholar 
van Dongen S, Abreu-Goodger C. Using MCL to extract clusters from networks. In: van Helden J, Toussaint A, Thieffry D, editors. Bacterial molecular networks: methods and protocol. New Jersey:Humana Totowa; 2012. p. 281–295.Bergeron A, Mixtacki J, Stoye J. A unifying view of genome rearrangements. In: Proceeding of the International Workshop on Algorithms in Bioinformatics. Berlin, Heidelberg: Springer; 2006. p. 163–73.Hilker R, Sickinger C, Pedersen CNS, Stoye J. UniMoG-a unifying framework for genomic distance calculation and sorting based on DCJ. Bioinformatics. 2012;28:2509–11.CAS 
PubMed 
PubMed Central 

Google Scholar 
Gelman A, Carlin JB, Stern HS, Dunson D, Vehtari A, Rubin DB. Bayesian data analysis: 3rd ed. Boca Raton:Chapman and Hall/CRC; 2013.Martiny AC, Ustick L, Garcia CA, Lomas MW. Genomic adaptation of marine phytoplankton populations regulates phosphate uptake. Limnol Oceanogr. 2020;65:S340–S50.CAS 

Google Scholar 
Garcia CA, Hagstrom GI, Larkin AA, Ustick LJ, Levin SA, Lomas MW, et al. Linking regional shifts in microbial genome adaptation with surface ocean biogeochemistry. Philos Trans R Soc Lond, B. 2020;375:20190254.CAS 

Google Scholar 
Ustick LJ, Larkin AA, Garcia CA, Garcia NS, Brock ML, Lee JA, et al. Metagenomic analysis reveals global-scale patterns of ocean nutrient limitation. Science. 2021;372:287–291.CAS 
PubMed 

Google Scholar 
Delmont TO, Eren AM. Linking pangenomes and metagenomes: the Prochlorococcus metapangenome. PeerJ 2018;6:e4320.PubMed 
PubMed Central 

Google Scholar 
Worden AZ, Binder BJ. Application of dilution experiments for measuring growth and mortality rates among Prochlorococcus and Synechococcus populations in oligotrophic environments. Aquat Microb Ecol. 2003;30:159–74.
Google Scholar 
Casey JR, Boiteau RM, Engqvist MKM, Finkel ZV, Li G, Liefer J, et al. Basin-scale biogeography of marine phytoplankton reflects cellular-scale optimization of metabolism and physiology. Sci Adv. 2022;8:eabl4930.CAS 
PubMed 
PubMed Central 

Google Scholar 
Berube PM, Rasmussen A, Braakman R, Stepanauskas R, Chisholm SW. Emergence of trait variability through the lens of nitrogen assimilation in Prochlorococcus. Elife. 2019;8:e41043.PubMed 
PubMed Central 

Google Scholar 
Luo HW, Benner R, Long RA, Hu JJ. Subcellular localization of marine bacterial alkaline phosphatases. Proc Natl Acad Sci USA. 2009;106:21219–23.CAS 
PubMed 
PubMed Central 

Google Scholar 
Schapiro JM, Libby SJ, Fang FC. Inhibition of bacterial DNA replication by zinc mobilization during nitrosative stress. Proc Natl Acad Sci USA. 2003;100:8496–501.CAS 
PubMed 
PubMed Central 

Google Scholar 
Hantke K. Bacterial zinc uptake and regulators. Curr Opin Microbiol. 2005;8:196–202.CAS 
PubMed 

Google Scholar 
Glass JB, Axler RP, Chandra S, Goldman CR. Molybdenum limitation of microbial nitrogen assimilation in aquatic ecosystems and pure cultures. Front Microbiol. 2012;3:331.CAS 
PubMed 
PubMed Central 

Google Scholar 
Rusch DB, Halpern AL, Sutton G, Heidelberg KB, Williamson S, Yooseph S, et al. The Sorcerer II Global Ocean Sampling expedition: northwest Atlantic through eastern tropical Pacific. PLoS Biol. 2007;5:e77.PubMed 
PubMed Central 

Google Scholar 
Noell SE, Barrell GE, Suffridge C, Morré J, Gable KP, Graff JR, et al. SAR11 cells rely on enzyme multifunctionality to transport and metabolize a range of polyamine compounds. 2021. https://www.biorxiv.org/content/10.1101/2021.05.13.444117v1.Binder BJ, Chisholm SW. Cell-cycle regulation in marine Synechcococcus sp strains. Appl Environ Microbiol. 1995;61:708–17.CAS 
PubMed 
PubMed Central 

Google Scholar 
Zinser ER, Lindell D, Johnson ZI, Futschik ME, Steglich C, Coleman ML, et al. Choreography of the transcriptome, photophysiology, and cell cycle of a minimal photoautotroph, Prochlorococcus. PLoS ONE. 2009;4:e5135.PubMed 
PubMed Central 

Google Scholar 
Hynes AM, Rhodes KL, Binder BJ. Assessing cell cycle-based methods of measuring Prochlorococcus division rates using an individual-based model. Limnol Oceanogr-Meth. 2015;13:640–50.
Google Scholar 
Vieira-Silva S, Rocha EP. The systemic imprint of growth and its uses in ecological (meta) genomics. PLoS Genet. 2010;6:e1000808.PubMed 
PubMed Central 

Google Scholar 
Weissman JL, Hou SW, Fuhrman JA. Estimating maximal microbial growth rates from cultures, metagenomes, and single cells via codon usage patterns. Proc Natl Acad Sci USA. 2021;118:e2016810118.CAS 
PubMed 
PubMed Central 

Google Scholar 
Chase AB, Karaoz U, Brodie EL, Gomez-Lunar Z, Martiny AC, Martiny JB. Microdiversity of an abundant terrestrial bacterium encompasses extensive variation in ecologically relevant traits. mBio. 2017;8:e01809–17.CAS 
PubMed 
PubMed Central 

Google Scholar 
Olm MR, Crits-Christoph A, Bouma-Gregson K, Firek BA, Morowitz MJ, Banfield JF. inStrain profiles population microdiversity from metagenomic data and sensitively detects shared microbial strains. Nat Biotechnol. 2021;39:727–36.CAS 
PubMed 
PubMed Central 

Google Scholar 
Flombaum P, Gallegos JL, Gordillo RA, Rincon J, Zabala LL, Jiao N, et al. Present and future global distributions of the marine Cyanobacteria Prochlorococcus and Synechococcus. Proc Natl Acad Sci USA. 2013;110:9824–9.CAS 
PubMed 
PubMed Central 

Google Scholar 
Brown MV, Lauro FM, DeMaere MZ, Muir L, Wilkins D, Thomas T, et al. Global biogeography of SAR11 marine bacteria. Mol Sys Biol. 2012;8:595.
Google Scholar  More

The current study is the first to directly compare the differences in eukaryotic community diversity by metabarcoding eRNA and eDNA from an estuarine benthic ecosystem. This is also the first study to compare the diversity of environmental RNA and DNA using two of the most common loci examined in metabarcoding applications: COI and 18S. Only a handful of studies have used eRNA to assess diversity through metabarcoding and have demonstrated some differences in detected diversity between eRNA and eDNA in metabarcoding19,29,38. Environmental DNA has many useful applications, but one of the downfalls of DNA is that it can be detected long after an organism has inhabited an area, such as from dead organisms and even significant distances away from the source1. While the persistence of eDNA may be useful in detecting the presence of endangered, rare, or invasive species in marine systems, the persistence of eDNA, such as legacy DNA, may skew the accurate representation of community structure immediately after a disturbance event or during the exact moment of sampling21. Environmental RNA proves to be a useful tool because it is far less persistent in marine environments2,27. In the current study, RNA provides a snapshot of living organisms present in the mesocosms at the time of sampling, compared to DNA which detects past and present organisms in the sample.Benthic ecosystems often have high biodiversity because they are dynamic environments with various substrates favorable for hosting communities39. To date, there is no one loci and associated primer pair that will effectively detect all eukaryotic organisms, and in particular metazoans, so it is beneficial to use multiple loci when looking for a variety of taxa6. Previous marine eDNA studies have demonstrated the preferred method of using two loci to achieve comprehensive metabarcoding for taxonomically diverse environments5,30,40. The current study further demonstrates the need to use at least two different loci for targeted PCR and sequencing to truly capture the wide diversity in rich systems, such as the top oxygenated layer of the marine benthos. Our results demonstrate the types of organisms detected using the 18S loci vary widely from the organisms detected by COI. The 18S loci also detected a higher number of organisms than using the COI loci. The 18S allowed greater detection of metazoans whereas COI was useful in the detection of Oomycota (i.e., a eukaryotic microorganism that resembles fungi) protozoa. Protozoa are often food sources for meiofauna41; therefore, using COI for protozoan detection and 18S for metazoan detection showcases multiple trophic levels present in the mesocosms of the current study. Utilizing multiple loci not only broadens taxonomic diversity detection but can also highlight multiple trophic levels for understanding food webs originating in benthic environments15.A study evaluating arctic benthic diversity similarly found that COI detected fewer taxa than 18S using eDNA metabarcoding, and there was only ~ 40% taxa overlap between markers at the class level42. In the current study, the COI marker using both nucleic acid templates yielded a higher percentage of unassigned taxa after filtering for presence in the majority of mesocosms compared to 18S. A possible explanation for the low metazoan detection by COI may be that the unassigned taxa are metazoans rather than more SAR organisms. Additionally, singly detected metazoans were filtered out of the analysis if they were not detected in at least 4 mesocosms. For example, one type of amphipod was detected in only 3 mesocosm by RNA using the COI marker, but was not detected in any mesocosms using DNA. Therefore, the amphipod observation was excluded from the COI results in Fig. 3 and Table S3 because it was not found in at least 4 mesocosms.Overall, RNA provides a broader assessment of benthic community structure than DNA, particularly when using two loci/ markers for sequencing. In the current study, we used nuclear 18S ribosomal RNA and DNA and mitochondrial COI RNA and DNA sequences as markers for metabarcoding. The number of copies of ribosomal RNA per cell is higher than the copies of ribosomal DNA, and the ratio of RNA: DNA is higher in single-cell organisms, such as protists43. The higher number of unique ASVs detected using eRNA is likely attributed to the higher number of RNA copies of each marker in small, single-cell organisms successfully amplified during PCR, therefore making rare organisms easier to detect. The DNA of highly abundant or higher biomass organisms may “drown out” (i.e., mask) the sequences from lower abundance and biomass organisms during PCR amplification, thus resulting in lower detected α-diversity. The increased detection of ciliates and protozoa using eRNA are consistent with other recent eRNA metabarcoding results that found higher ciliate and protozoan diversity compared to eDNA using the same Uni18S primers19. Positive correlations between organism biomass and sequence copy numbers have been demonstrated for DNA metabarcoding conducted on invertebrate species44. In the current study, RNA allowed for the detected of both larger meiofauna and smaller microfauna, which is optimal for assessing true biodiversity with molecular assays. Chaetonotida, a type of gastrotrich, was the only taxa detected in 4 of the mesocosms using DNA that was not found with RNA. It is possible the chaetonotida may have died during the experiment due to sensitivities to new environmental conditions, and therefore were not detected with RNA. The temperature of the flowing seawater in the mesocosms was approximately 18 °C, and many marine chaetonotida prefer 23–28 °C and high organic matter substrate45.Previous studies have compared the accuracy of conventional morphological identification to molecular metabarcoding methods for assessing biodiversity. In estuarine-specific studies, metabarcoding methods are able to detect the majority of taxa identified with traditional methods and often detected higher species richness, or higher numbers of unique organisms, that was not found conventionally10,46,47. The limiting factor for higher resolution of taxonomic identification is the availability of species-specific sequences in barcoding databases48. However, barcoding databases are becoming more robust as an increasing number of researchers contribute high quality sequencing data to databases7. Therefore, eRNA metabarcoding techniques perform similarly to conventional morphological methods, and may even uncover higher biodiversity in systems like estuaries where meiofauna have been historically understudied and identified.Although eRNA α-diversity is higher compared to eDNA, there is some overlap between ASVs detected with eRNA and eDNA. The higher percentage of overlap between eDNA and eRNA ASVs is predominantly seen in the ASVs detected in all 7 of the mesocosms. The increased overlap in ASVs detected in all 7 mesocosms compared to the relatively lower overlap of ASVs found in only 1 of the 7 mesocosms is due to the filtering of random organisms found in only 1 mesocosm. Detection of a unique organism in a single mesocosm is likely not representative of the sample community and filters potential artifacts that may be introduced during cDNA synthesis from eRNA. However, all uniquely identified ASVs are used in the β-diversity analysis, so the higher number of unique ASVs detected from eRNA are likely driving the significant differences observed between eRNA and eDNA β-diversity. It is possible that using eRNA could increase the statistical power of a study design compared to eDNA because eRNA detects a higher number of ASVs. It is unlikely that a higher number of RNA ASVs could be due to splice variants contributing to unique sequences because the amplified regions of both markers do not contain introns. Therefore, no splicing of transcripts would be expected. Thus, the detected DNA ASVs are from living organisms in the mesocosms and the higher diversity of ASVs detected from RNA demonstrate that RNA is a more suitable option for assessing diversity of living organisms.It is evident that the mesocosms in the current study were rich with meio- and microfauna due to the number of unique organisms and broad diversity of different taxa. It is likely that collecting samples for nucleic acid extraction directly from the field site may result in higher diversity because there is no mechanical disturbance during the laboratory acclimation period and the presence of other organisms in the system, such as fish, macroinvertebrates, or birds. eDNA molecular abundance in samples has been shown to correlate to actual organismal abundance in laboratory environments, but the same correlation is not as apparent in field samples, which is likely due to a variety of collection and processing methods49. eRNA may be the better nucleic acid template for field collection once flash frozen, especially for the detection of protozoans. Future studies will compare the diversity of eukaryotic communities detected using eDNA and eRNA collected directly from the field rather than from sediment core mesocosms. Repeated sampling from a field site may help reduce transient eRNA detection when establishing accurate baseline community composition in field-based biomonitoring studies.Recently, meiofaunal organisms and communities are being explored as bioindicators, which are organisms whose presence are indicators for environmental stress and pollution17. For example, lower meiofaunal diversity and abundance is associated with higher pollution in harbors, and the presence of some genera of nematodes are correlated with higher concentrations of polycyclic aromatic hydrocarbons because they are more tolerant to pollution50. A previous study that used field-collected mesocosms from the same location as our current study found similar phyla detected with a metabarcoding approach; the majority of the sequences detected in benthic communities were from nematodes, arthropods, and the microfaunal SAR clade10. Similarly, the majority of the ASVs detected from the current study also corresponded with nematodes, arthropods, and the SAR clade, as well as other commonly detected meiobenthic organisms, such as polychaetes and Homalorhagida (i.e., mud dragons). A recent study exposed a benthic foraminiferal community to chromium, and found that eRNA metabarcoding was more robust for detecting changes in diversity at lower chromium concentrations compared to eDNA51. The eRNA metabarcoding method used in the current study detected meiofaunal taxa typical of marine or estuarine environments. Therefore, eRNA metabarcoding may be useful for efficiently identifying bioindicator species or taxa impacted by exposures to different contaminants and environmental stressors to aid with management of aquatic systems. Another advantage of eRNA is that RNA provides functional information about how organisms response to stress through altered transcription of activated pathways. eRNA will likely be a more powerful tool than eDNA because it allows for the detection of both bioindicator species and, in the future with increase development of genetic databases, environmental detection of biomarkers of stress through increased transcription of response genes.Many academic researchers are adopting molecular methods using High Throughput Sequencing as the future of biomonitoring surveys; however, few regulatory and environmental management organizations/ agencies have adopted metabarcoding into their routine biomonitoring practices for regulatory purposes. In marine benthic communities, metabarcoding provides a comprehensive assessment of diversity and is useful for detecting a broader array of organisms in biomonitoring surveys especially when using two markers47. eDNA is also useful for monitoring discrete communities (i.e., benthic versus pelagic)52, so it is possible that using eRNA could provide vital information about living organisms in specific environments compared to eDNA. Metabarcoding sequencing and bioinformatic approaches for benthic environments vary among studies, thus requiring some standardization between methods to further advance the use of metabarcoding in conservation and regulatory applications30,53.Like eDNA, some advancements must be made with eRNA to be used as a quantitative tool in molecular ecology. Validation of eDNA metabarcoding for assessing relative abundance of species is rapidly progressing by correlating laboratory studies of DNA shedding with field experiments1,54. Similar validation techniques can be used to develop eRNA metabarcoding as a quantitative or semi-quantitative method. There is growing interest in optimizing eRNA extraction protocols from different types of environmental media to standardize the use of eRNA in downstream molecular applications55. Thus, standardizing eRNA protocols will help with integration into environmental management toolkits for regulatory purposes.RNA poses unique barcoding challenges compared to DNA because the number of RNA transcripts from a gene are not always present in the same proportion compared to the gene copy number per genome (i.e., one DNA copy per cell), especially for differentially expressed genes. However, one possible way to work around this issue is to utilize constitutively expressed marker loci where transcription is generally stable and unaffected by environmental stressors, such as those used as reference genes for quantitative real-time PCR49. Fortunately, many loci chosen for metabarcoding purposes fit this criterium; the transcription of 18S and COI remain steady within the cell regardless of environmental stress.Collecting sediment cores from the environment and bringing them into controlled laboratory settings for community analysis through eRNA metabarcoding is a powerful tool that opens opportunities for this method to be used in a broad range of fields. For instance, field-collected mesocosms could be used in controlled settings to investigate the effects of individual or mixtures of toxicants on entire community and population-level outcomes. Marine sediments are often the ultimate sink for environmental contaminants, such as organic pollutants , heavy metals56, and plastic particles57, yet few studies investigate actual community-level changes in contaminant exposures. Marine benthic environments have high biodiversity, but the breadth of diversity in micro- and meiofaunal organisms is often understudied because traditional morphological methods are immensely time consuming. Sediment core mesocosms could also be used to understand the effects of global climate change stressors, such as fluctuating temperatures, surface water salinity and pH, on communities as well as conventional and emerging contaminants in combination with climate change stressors. Additionally, this method could also be used to understand how communities respond to a significant disturbance event or smaller series of stressors, which are often difficult to measure in environmental settings32,58. In these applications, eRNA is favorable for constructing community composition at a specific moment of the experiment to better regulate anthropogenic causes of environmental stress. More

Electrohydrogenesis effluent as a potential biofertilizerTo characterize the electrohydrogenesis left-over residues as potential biofertilizers, the sample from the operating reactors was performed a 16S rRNA sequencing test, and interestingly, the results revealed that the bio-electrohydrogenesis effluent was enriched with various microorganisms including plant growth-promoting microbes that display biofertilizer-like features. Among the well-known plant-promoting bacterial genera observed in DF-MEC residues included Azospirillum, Mycobacterium, Chryseobacterium, Paenibacillus, Rhizobacter, Pseudomonas, Achromobacter, Bradyrhizobium, Actinomyces, Sphingomonas, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium, Gordonia, Rhodococcus, Bacillus, Methylobacterium-Methylorubrum, Microbacterium, Flavobacterium, Devosia, Acinetobacter, Mesorhizobium, Enterobacter, Aeromonas, Beijerinckia, etc.24,25,26 (Fig. 2). Lots of investigations working on the feasibility of using biofertilizers other than chemical fertilizers have revealed that those aforementioned microbes play a major role in providing the required nutrients for enhanced crop yield.Figure 2The abundance of the Plant growth-promoting bacteria (genus level) detected from the DF-MEC digestate (%).Full size imageNitrogen-fixing microorganismsThe detected nitrogen-fixing microorganisms from the electrohydrogenesis effluent include Azospirillum sp. (0.11 ± 0.02%), rhizobia (Rhizobium (0.058 ± 0.02%), Bradyrhizobium (0.11 ± 0.04%), and Mesorhizobium (0.1 ± 0.03%)), and Beijerinckia (0.08 ± 0.03%) (Fig. 2) and were repeatedly reported for their superior contribution to the plants’ nitrogen requirements through biological nitrogen fixation, which is an important component of sustainable agriculture25. Although the atmosphere counts about 78% N2, it couldn’t be used by plants in its natural state. Prior to getting used by plants, it needs to be converted to ammonia, which is the readily assimilable form of nitrogen by plants/or crops via a biological nitrogen fixation mechanism25. The biological Nitrogen fixation mechanism is summarized in Fig. 3.Figure 3Mechanism of nitrogen fixation bio-catalyzed by nitrogenase enzyme. The plant growth-promoting bacteria produce nitrogenase which is a complex enzyme consisting of dinitrogenase reductase and dinitrogenase. This complex enzyme plays a major role in molecular N2 fixation. Dinitrogenase reductase provides electrons and dinitrogenase uses those electrons to reduce N2 to NH3. However, oxygen is a potential threat to this process since it has the ability to get bound to the enzyme complex and make it inactive and consequently inhibit the process. Interestingly, bacterial leghemoglobin has a strong affinity for O2 and thus gets bound to free oxygen more strongly and effectively to suppress the available oxygen effects on the whole process of nitrogen fixation.Full size imagePhosphate-solubilizing microorganismsFurthermore, various phosphate-solubilizing and mineralizing strains were also found in bioelectrohydrogenesis residues collected from our DF-MEC integrated reactors. Among those microorganisms with the ability to solubilize/metabolize the insoluble inorganic phosphorus, the dominant bacterial genera included Pseudomonas (0.65 ± 0.15%), Bacillus (0.44 ± 0.11%), Rhodococcus (0.04 ± 0.009%), Rhizobium (0.05 ± 0.02%), Microbacterium sp. (0.04 ± 0.01%), Achromobacter (0.16 ± 0.07%), and Flavobacterium (0.058 ± 0.014%) (Fig. 2). Though enormous amounts of phosphorus are available in the soil, its high portion never contributes to plant growth in its primitive state, unless it is bio-transformed into absorbable forms including monobasic and dibasic. Microbial phosphate solubilizing mechanisms are well described in Fig. 4.Figure 4Inorganic phosphorus solubilization by phosphate-solubilizing rhizobacteria. A bacterium solubilizes inorganic phosphorus through the action of low molecular weight organic acids such as gluconic and citric acids. The hydroxyl (OH) and carboxyl (COOH) groups of these acids chelate the cations bound to phosphate and thus convert insoluble phosphorus into a soluble organic form. The mineralization of soluble phosphorus occurs by synthesizing different phosphatases which catalyze the hydrolysis process. When plants incorporate these solubilized and mineralized phosphorus molecules, eventually, overall plant growth and crop yield significantly increase.Full size imagePhytohormone-producing microorganismsIn this current work, the electrohydrogenesis effluent also contained bacterial genera such as Mycobacterium (0.77 ± 0.18%), Allorhizobium (0.05 ± 0.02%), Pararhizobium (0.05 ± 0.02%), Paenibacillus (1.18 ± 0.24%), Bradyrhizobium (0.11 ± 0.04%), Rhizobium (0.05 ± 0.02%), Acinetobacter (0.14 ± 0.02%), and Azospirillum (0.11 ± 0.025%) (Fig. 2) that have the ability to synthesize indole-3-acetic acid/indole acetic acid (IAA) through indole-3-pyruvic acid and indole-3-acetic aldehyde25. IAA is a well-known type of phytohormone that enhances plant/crop growth. Particularly, Azospirillum sp., also produce various phytohormones namely cytokinins, gibberellins, ethylene, abscisic acid and salicylic acid, auxins, vitamins such as niacin, pantothenic acid, and thiamine. The conceptional model delineating the positive effects of inoculation with Azospirillum sp. a phytohormones-producer plant growth-promoting rhizobacteria and its detailed functions on plant growth are summarized and illustrated in Fig. S1. Therefore, the existence of those rhizobacteria in the bioelectrohydrogenesis residues further implies the suitability of considering the DF-MEC left-over residues as potential biofertilizers.Heavy metals-bioremediating microorganismsSome other bacterial genera with the ability to bioremediate the heavy metal toxicity were also found within the bioelectrohydrogenesis left-over residues as well. Among the detected plant growth-promoting bacterial genera; Rhizobium (0.058 ± 0.023%), Mesorhizobium (0.1 ± 0.026%), Bradyrhizobium (0.11 ± 0.04%), Pseudomonas (0.65 ± 0.15%), and Achromobacter (0.16 ± 0.077%) were reported for their key contribution to alleviate the toxicity of the heavy metals via bioremediation process and improve the soil quality for a relief plant development26 (Fig. 2). Other detected heavy metals-bioremediating microorganisms’ species were Chryseobacterium sp. (0.08 ± 0.007%), Azospirillum (0.11 ± 0.02%), Bacillus (0.44 ± 0.11%), Enterobacter (8.57 ± 0.9%), Gordonia (0.06 ± 0.02%), Paenibacillus (1.18 ± 0.24%), Pseudomonas (0.65 ± 0.15%), and Actinomycetes (0.36 ± 0.05%) that either use microbial siderophores or enzymatic biodegradation process.Electrohydrogenesis left-over residues as a potential source of essential elements for plant growthAs aforementioned in “Materials and methods” section, the electrohydrogenesis left-over residues contained diverse microbial communities that degraded the MEC substrate and generate biogas and inorganic compounds. Moreover, it has been reported that those inorganic nutrients are generally available in fermentation effluent in readily plant-utilizable formats owing to substrate mineralization27. Beside detecting various plant growth-promoting microorganisms in the electrohydrogenesis effluent, a larger number of mineral elements essential for promoted growth and development of crop plants were also investigated and analyzed from the residues. The detected primary and secondary macro-elements’ concentrations in the residues were arranged in decreasing order as follows P  > S  > Na  > K  > N  > Ca  > Mg. Interestingly the findings show that the residues abundantly contained Phosphorus (2.766 × 103 mg/L), Nitrogen (274 mg/L), Potassium (282 mg/L), Calcium (17.66 mg/L), Magnesium (16.3 mg/L), Sulfur (1.225 × 103 mg/L), and Sodium (294.3 mg/L) which are well known as macro-nutrients needed in larger amounts for enhanced plant/ crop growth (Fig. 5).Figure 5Macro-, and micronutrients detected from the bio-electrohydrogenesis left-over residues (mg/L).Full size imageMoreover, small amounts of the microelements including Ni, Pb, Zn, Cu, Cr, Hg, Cd were also found in the electrohydrogenesis residues, and consistently these elements are generally required in small quantities for the development of plants (Fig. 5), otherwise, their high concentrations are toxic for the plant cells thus suppress or inhibit plant growth. The detected concentrations for the main microelements in this current research ranged only from 0.36 to 9.6 × 10–5 mg/L and were all reported to play fundamental roles in plant metabolic reactions.Cultivation of the leguminous crops using electrohydrogenesis left-over residues as fertilizerAfter evaluating the plant-growth promoting bacterial communities and the macro- and micronutrients required for plant/crop growth in the electrohydrogenesis left-over residues, the latter was directly used as fertilizer to grow three different plant species including tomato, chili, and brinjal as afore-described in the “Materials and methods” section. To access the potentials of the electrohydrogenesis effluent as fertilizer, the plants grown in the soil amended with the effluent (Soil + Effluent), were directly compared with their corresponding control plants (Soil + water). The results indicated that at the end of 1st month, the plants with effluent grew faster and generated a good amount of branching than the control plants (see Fig. 6), possibly due to the availability of both microbial species with bio-fertilizing aspects and micro-and macronutrients in the effluent.Figure 6Analysis of the plant growth at the end of the 1st month of cultivation. (a) Tomato in soil with effluent, and its control without effluent (b); (c) Chilli grown in soil with effluent, and its control without effluent (d); and (e) brinjal grown in soil with effluent, and its corresponding control grown without effluent (f) (after 2 months).Full size imageFor instance, tomato (Solanum lycopersicum L.) and chilli (Capsicum annuum L.) height in soil + electrohydrogenesis effluent was ~ 36.9 ± 2.1 cm and ~ 32.6 ± 0.8 cm respectively which was ~ 2.03 and ~ 1.2 times the height of their corresponding plant species in the control protocol, respectively (see Fig. 7). However, the brinjal species (Solanum melongena L.) didn’t show any remarkable height differences in both protocols after a month of cultivation (data not shown), probably due to their low adaptative characteristics to the new environment. However, after the 2nd month, the brinjal height in soil + effluent became 2.7 times that of the brinjal control cultivated without effluent (see Fig. 6e,f). Moreover, both the number of the plants’ leaves and their length in plants cultivated in soil + effluent, were remarkably higher than in plants grown without the supply of the effluent.Figure 7Daily plant growth analysis within one month of cultivation. (a) Tomato growth monitoring, (b) Chili growth analysis.Full size imageAt the end of the 3rd month, the plants in soil + electrohydrogenesis effluent generated more fruit with big size than the control plants (see Fig. 8), but the tomato (Solanum lycopersicum L.) didn’t generate fruits in both protocols at that time probably due to the high weather temperature that inhibitory affected its continuous growth, as previously reported that tomato species are generally so sensitive to temperature change28,29. The final yield was evaluated in terms of the size and number of fruits per cultivated plant. Chili cultivated in soil with MEC effluent generated 3 fruits/plant and its corresponding control without effluent produced only 1 fruit/plant. The chili fruit size in soil + effluent was 16 cm, approximately 18.7% higher than its corresponding control. Moreover, at the time of collecting data, the brinjal plant cultivated in soil with MEC effluent generated brinjal fruits whereas its corresponding cultivated without electrohydrogenesis effluent started flowering (see Fig. 8). These further indicate the significant contribution of the electrohydrogenesis effluent in speeding up the plant growth. Herein, the electrohydrogenesis left-over residues have notably improved the soil quality and significantly promoted the plants’ phenology characterized by plant growth, the generation of new leaves, flowering, and the production of fruits.Figure 8Analysis of plant growth characterized by the flowering and fruiting process at the end of the 3 months. (a) Chili grown in soil with effluent, and its control without effluent (b); (c) brinjal grown in soil with effluent and its corresponding control grown without effluent (d).Full size image More

We applied a modified version of a planning framework for CWR conservation25,26 which has been used by numerous countries of Europee.g.29,63,64, Americae.g.65, Africa30 and Asia66,67. We addressed the following main steps of the toolkit (see Spanish version49): (i) CWR checklist, i.e., creating a list of CWR taxa distributed in an area (Supplementary Data 1), (ii) CWR inventory, i.e., taxa selection and collation of ancillary data, including taxonomic data (Supplementary Data 2), (iii) taxa extinction risk assessment (Table 1, Supplementary Data 3), and (iv) a systematic conservation planning assessment, i.e., spatial analyses to assess conservation areas (Fig. 1). We only provide a brief description of steps i-iii, as these are thoroughly described in Goettsch et al.2. Here, we focus on the systematic conservation planning assessment, introducing an approach in order to identify conservation areas for CWR that account for genetic differentiation in a spatially explicit way, through the use of proxies of genetic differentiation (Fig. 1).During the process -framed under the project “Safeguarding Mesoamerican crop wild relatives” (https://www.darwininitiative.org.uk/project/23007/)- more than 100 experts from academic, governmental, and non-governmental organizations from El Salvador, Guatemala, Honduras, Mexico, the UK, and IUCN participated in six workshops, shared data, and provided fundamental knowledge and feedback at each project stage to ensure accurate, reliable and robust information for next steps. The checklist, inventory and risk assessment were collaboratively developed between partners of El Salvador, Guatemala, and Mexico (hereafter, Mesoamerica; Goettsch et al.2). The spatial analysis to identify areas for in situ and ex situ conservation of CWR was done independently by each country.To assess conservation areas of CWR in Mexico, we developed proxies of genetic differentiation that account for evolutionary processes by including historical and environmental drivers of genetic diversity (see the Methods section ‘Proxies of genetic differentiation’). In addition, we used criteria such as information on taxon-specific tolerance to human-modified habitats and IUCN extinction risk category. We applied a systematic conservation planning approach and performed spatial analysis using the software Zonation50. We compared different scenarios to represent genetic diversity of CWR based on potential species distribution models (SDM) and proxies of genetic differentiation.Study areaMesoamerica is a cultural region encompassing the territories of Belize, Guatemala, El Salvador, the southern part of Mexico and parts of Honduras, Nicaragua and Costa Ricasee 2. In this study, we also included the dry areas of northern Mexico that are part of Aridamerica68 and the Nearctic biogeographic realm69 to account for the full extent of the geographic range of many taxa included in the extinction risk assessment2.For the assessment of conservation areas, we focused on Mexico, which is one of the most biodiverse countries in the world70. The Mexican territory covers 80% of the landscapes of the region called Mesoamerica. Its high biological diversity is attributed to its geographic, topographic, climatic, geological and cultural characteristics, which, among other factors, shaped the distribution of an extraordinary variety of ecosystems and species with high levels of endemism and species turnover among different regions32,71,72,73. In particular, the high genetic variation within populations of landraces and CWR is the result of past and ongoing sociocultural processes occurring in a wide range of distinct environmental conditions74,75.(i) CWR checklist and (ii) CWR inventoryThe compiled CWR checklist included ~3000 species and subspecies of 92 genera and 45 families of plants that belong to the same genus of a crop cultivated in Mesoamerica, or wild plant collected for food or other uses in the region (Supplementary Data 1).The first set of criteria were established in preparation for the first stakeholder workshop. The following criteria were applied at the genus level to compile the CWR inventory: (1) occurrence of wild relatives of cultivated plants or crops that were domesticated in Mesoamerica; (2) existence of research groups working on taxa that could support the extinction risk assessment; and (3) relation to a crop of economic and nutritional importance at local, national and regional levels, or cultivars known to require genetic improvement.To narrow the list for the inventory and extinction risk assessment, similar criteria were agreed upon in the same workshop and applied at the species level: (1) native distribution in Mesoamerica, incl. Aridamerica; (2) related to a crop of economic or social importance based on production and nutritional value; (3) related to a taxon for which Mesoamerica is the center of origin or domestication; (4) constitutes part of the primary or secondary gene pool, and in some cases the tertiary gene pool76. The primary gene pool consists of wild plants of the same species as the crop and thus their mating produces strong fertile progeny. The secondary gene pool is composed of wild relatives distinct from cultivated species but closely related as to produce some fertile offspring (same taxonomic series or section in the absence of crossing and genetic diversity information, see the ‘taxon group’ concept proposed by Maxted and collaborators77, Supplementary Note 5). The tertiary gene pool (same subgenus in the taxon group concept) corresponds to CWR that are more distant relatives to the taxa of the primary gene pool, but can have important adaptive traits which can be used with specific breeding techniques. This provided a preliminary list of 514 CWR taxa related to avocado, cotton, amaranth, cocoa, squash, sweet potato, chayote, chili pepper, cempasuchil, bean, sunflower, maize, papaya, potato, vanilla, and yuca (Supplementary Data 2).The list had to be further reduced due to time and funding restrictions to include those genera which when added together would include no more than 250 taxa, and that the taxonomic groups could be comprehensively assessed and their taxa evaluated throughout their entire range. Thus, not all species in the group necessarily met the criteria previously mentioned. See the final Mesoamerican CWR inventory in Supplementary Data 3; see summary in Table 1.(iii) Taxa extinction risk assessmentFull methodological details and results of this section are described in Goettsch et al.2. Summarizing, during the process 224 taxa were evaluated according to the International Union for Conservation of Nature, IUCN, Red List Categories and Criteria78. The IUCN Red List is a critical indicator to identify species most vulnerable to extinction considering a set of criteria, i.e., species’ population trends, size, structure, and geographic ranges. A Red List workshop with the participation of 25 experts from different project partner institutions and IUCN specialists was organized to assess the extinction risk of taxa. The threat analysis included not only species, but subspecies and subpopulations (i.e. races) for some groups (Supplementary Data 3, see summary in Table 1).(iv) Systematic conservation planning assessmentTo undertake the following spatial analyses we focused on the dataset of 224 CWR described above, which is representative of the CWR of the main crops of Mesoamerica (10 genera, Table 1).Species distribution modelingTo compile occurrence records, hundreds of data sources were consulted, including published and personal databases of the project participantse.g.79,80,81,82, the Agrobiodiversity Atlas of Guatemala (https://www.ars.usda.gov/northeast-area/beltsville-md-barc/beltsville-agricultural-research-center/national-germplasm-resources-laboratory/docs/atlas-of-guatemalan-crop-wild-relatives), the Global Biodiversity Information Facility (GBIF, https://www.gbif.org/), and Mexico’s Biodiversity Information System (SNIB, http://snib.mx/).To generate potential species distribution models (SDM), we used more than 13,000 occurrence records (Supplementary Data 4), that were standardized and curated by experts to generate the range maps of taxa as part of the extinction risk assessment, which were published in IUCN Red List (https://www.iucn.org/news/species/202109/threats-crop-wild-relatives-compromising-food-security-and-livelihoods). Spatial resolution of the SDM was 1 km2. SDM were obtained for taxa with more than 20 unique occurrence data in a 1 km2 grid covering the study extent to reduce uncertainty when using smaller sample sizes83. We used 19 bioclimatic variables and other climatic variables, such as annual potential evapotranspiration, aridity index, annual radiation, slope, and altitude84,85,86. Climate data represents annual and seasonal patterns of climate between 1950 and 2000. Also, we used a variable that described the percentage of bare soil and cultivated areas87. Collinearity between variables was assessed with the ‘corselect’ function of the package fuzzySim version 1.088, using a value of 0.8 and the variance inflation factors as criteria to exclude highly correlated variables.We used MaxEnt version 3.3.1, a machine-learning algorithm that uses the maximum entropy principle to identify a target probability distribution, subject to a set of constraints related to the occurrence records and environmental data89,90. Model calibration area for each taxon included those ecoregions where the taxon has been recorded; we used the terrestrial ecoregions dataset69. We did this based on the calibration area or ‘M element’ of the BAM diagram that refers to areas that have been accessible to the taxon via dispersal over relevant periods of time91,92. We randomly sampled 10,000 background localities from the selected areas.To reduce model complexity without compromising model performance, we built several models by varying the feature classes (FC) and regularization multipliers (RM) (see refs. 93,94,95) using R 3.6.096 and ‘ENMeval’ version 0.3.0 package97. FC determines the flexibility of the modeled response to the predictor variables, while the RM penalizes model complexity93. Occurrence records were randomly divided into 70% for model selection, and 30% of data was withheld for model validation. ENMeval carries out an internal partition of localities to test each combination of settings. Therefore, we selected the random k-fold method to divide localities into four bins. We build models with six FC combinations and varied RM values ranging from 0.5 to 4.0 in 0.5 increments. Optimal models were selected using Akaike’s Information Criterion corrected for small sample sizes (⍙AICc = 0). This method penalizes overly complex models and helps to choose those with an optimal number of parameters. However, it has been shown that the number of model parameters may not correctly estimate degrees of freedom98, and that model selection should not be selected solely with one measure99. Thus, we used 30% of the withheld data to test the area under the curve (AUC) of the receiver operating characteristic, and the omission error under a 10 percentile training threshold.We used the ten percentile or minimum training presence threshold to obtain binary maps of the presence and absence of suitable areas for species distribution. We asked experts of each taxonomic group who were also involved in the extinction risk assessment to select one of these two options and to indicate possible overestimated areas, which were then eliminated case by case using the information of Mexican ecoregions100 and watersheds101. Eight models were binarized with the minimum training presence threshold; for the other models we used the 10 percentile threshold. See MaxEnt performance and significance of SDM at Supplementary Data 5. AUC values ranged from 0 to 1; 0.5 indicated a model performance not better than random, while values closer to 1 indicated a better model performance; here we used SDM showing AUC values higher than 0.7. For Phaseolus and Zea, we used SDM that were previously generated by Delgado-Salinas et al.102, and Sánchez González et al.103, respectively. SDM for 116 taxa were validated by experts of each taxonomic group. See references and download links at Supplementary Data 6.For the conservation planning analysis of Mexico, we clipped the models to the Mexican territory, and trimmed the continuous SDM using the binary SDM to keep pixel values of areas with elevated probability of taxa presence. For taxa without SDM, we included the occurrence records of these taxa in the spatial analysis by using the information on observation location, i.e., coordinates (see Supplementary Data 3). This is done by enabling the function ‘species of special interest’ (SSI). See further details in the method section ‘Final conservation analysis’.Proxies of genetic differentiationTo identify proxies of genetic differentiation in an explicit, efficient, and repeatable way, we included environmental and historical drivers of genetic diversity. For this, we first divided Mexico into 27 Holdridge life zones (Supplementary Fig. 2, Supplementary Data 8), which we then subdivided according to phylogeographic studies that have found genetic differentiation among populations of several taxa (see division of each life zone into proxies in Supplementary Fig. 4; Supplementary Fig. 3 provides a general geographical overview of Mexico and main geographic references mentioned in Supplementary Fig. 4). The literature review was done searching for the words “phylogeography” and one of the following: (i) name of the Mexican biogeographic zones, (ii) “Mexico” + an ecosystem name (e.g. “Mexico” “rainforest”) or (iii) “Mexico” + lowlands/highlands. See list of references used in this study in Supplementary Data 9.In addition, we manually reviewed the citations to the most cited papers of the previous search. Reviews and meta-analyses were also included, although we excluded studies performed in CWR to show that our approach can be used without prior information on this group. As more studies on such taxa become available, they can be used to fine-tune the proxies of genetic differentiation. We focused on terrestrial species including plants, animals, and fungi (Supplementary Data 10) except to subdivide a life zone covering the coasts of the California Peninsula, where we could not find studies on terrestrial taxa so we included studies on fish species (see Supplementary Fig. 4).Since most of the life zones cover large territories, and complete phylogeographic congruence among different taxa is uncommon, we targeted to represent general trends that would likely occur across diverse species, instead of trying to represent fine idiosyncratic patterns of genetic differentiation. For instance, although distribution ranges of highland taxa shifted during the Pleistocene climate fluctuations, in general populations persisted (glacial-interglacial periods) within the main mountain ranges, while lowland populations were ephemeral (only glacial periods). So, gene flow among mountain ranges was more limited than within them. As a result, genetic differentiation among mountain ranges of different biogeographic provinces has been widely documented32, so we used this general pattern to subdivide the life zones that occur in highlands. These types of patterns are particularly relevant for a country like Mexico, due to its complex topography, tropical latitude, and geographic features of different ages, which promote population differentiation among the Mexican main geographic features. To translate the phylogeographic information into a spatial context, we used biogeographic regions, basins, topographic or edaphic data to split the life zones into different subzones using the best fitting cartography to represent the phylogeographic patterns (Supplementary Fig. 4).We obtained 102 proxies of genetic differentiation for Mexico (Supplementary Fig. 5). We validated our findings by using available genomic data of an empirical study of a wild relative of maize, the teosinte Zea mays subsp. parviglumis, which was not included in the literature review in order to test the usefulness of our approach regarding the lack of genetic data. The dataset includes ca. 1800 occurrence records and ca. 30,000 SNPs48. Sampling localities were not used for distribution modeling. Admixture groups per population were estimated for K1 to 60. According to the population analysis, Z. mays subsp. parviglumis is structured in 13 genetic clusters along a longitudinal gradient (Fig. 3a–c). We used the K = 13 for plotting based on the Cross-Validation error. The proportion of each genetic cluster was estimated by sampling locality and plotted using pie charts over the map (Supplementary Fig. 6). Then, using the data layer of the SDM subdivided by proxies of genetic differentiation, we extracted which was the proxy most frequent in a 5 km buffer for each sampling locality. The Admixture plot was ordered by all genetic clusters and subdivided by the proxy of genetic differentiation most frequent for each locality. In addition, we calculated a principal component analysis (PCA) and projected into a score plot the first three components. Individual samples were colored by the proxies where they fell in the 5 km buffer (Fig. 3c). To compare how genetic variation was represented by the different scenarios we plotted the proportion of the area of each proxy as given by the potential SDM according to two different scenarios (only considering SDM; combining SDM*PGD) considering 20% of Mexico’s terrestrial area (Fig. 3d). Analyses were run in R version 3.5.196 using the R packages pcadapt version 4.3.3104, ggplot2 version 2_3.3.3105, readr version 1.4.0106, gridExtra version 2.3107, ggnewscale version 0.4.5108, scatterpie version 0.1.5109, pophelper version 2.3.1110, raster version 3.4-5111, rgdal version 1.4-8112, rgl version 0.107.10113, and sp version 1.4-4114,115.Habitat preferenceWe considered habitat preference to refine the presence of CWR in the planning process; thus minimizing commission errors and highlighting areas that more probably contain taxa116. For each taxon, experts assessed its habitat preference (1: high preference; 0.5: low preference; 0.1: no preference) according to the following categories: (i) well-conserved vegetation (i.e. primary vegetation), (ii) human-impacted vegetation (i.e. secondary vegetation), (iii) less intensive rainfed and moisture agriculture, (iv) intensive rainfed and moisture agriculture, (v) irrigated agriculture, (vi) induced and cultivated grasslands and forests, and vii) urban areas (Supplementary Data 11). To spatially delimit these classes, we used the land use cover and vegetation map for Mexico117, and assessed seven main categories of land cover by grouping the map legend (Supplementary Fig. 9). To differentiate between less intensive and intensive cultivated areas, we followed Bellon et al.56, who associated the presence of native maize varieties of Mexico to occur in municipalities with average yields of less than or equal to 3 t ha-1 using agricultural production data from 2010 from the Information System of Agrifood and Fisheries (SIAP), and selected the municipalities with the established average maize yield. We combined the municipality layer with the land cover map to differentiate areas of high and low agricultural intensity. To generate taxon-specific habitat layers, we associated the habitat preference classes established by experts to the land cover map aggregated into seven major land cover categories, using R 3.6.096 and the following packages: raster version 3.4-5111 and rgdal version 1.4-8112. We obtained habitat maps for 116 taxa with SDM.Preliminary analysisWe generated five preliminary scenarios to explore different approaches to include conservation features for maximizing the representation of intraspecific diversity as given by taxa and proxies of genetic differentiation, i.e., representation of proxies within a taxa range (Supplementary Fig. 7): (i) “SDM” scenario, included 116 SDM, which we used as base scenario to examine the representation of taxa and proxies of genetic variability (n = 116); (ii) “SDM + LZ” scenario, included 116 SDM and 27 layers representing Holdridge life zones to consider environmental variation (n = 143); (iii) “SDM + PGD” scenario, included 116 SDM and 102 layers representing each proxy of genetic differentiation individually (n = 218); (iv) “SDM*PGD” scenario, included 5004 input layers representing the intersection of SDM and PGD (n = 5004; combining 116 SDM with 102 proxies resulted in 11,832 layers, but as some of the intersections produced empty outputs given the extension of SDM that do not cover all Mexico, for further analysis we used 5004 input layers with value data. To subdivide the layers, we used ArcGIS version 10.2.2118; to filter the layers, we used R 3.5.196.); (v) “SDM and PGD as ADMU” scenario, included 116 SDM as the main conservation features, while integrating one single layer of proxies of genetic differentiation to consider each of them as planning units by using the ‘Administrative units’ function. Analysis was done in Zonation50,119.We compared the results by assessing 20% of Mexico’s terrestrial area (Fig. 5b) to perform statistical analysis in R 3.5.196 using the following packages: purrr version 0.3.4120, ‘dplyr’ version 1.0.2121, ‘ggplot2’ version 2_3.3.3105, ‘raster’ version 3.4-5111, ‘scales’ version 1.2.0122, ‘sp’ version 1.4-4114,115, ‘tidyr’ version 1.0.2123, and ‘vegan’ version 2.6-2124. The area threshold was established based on Aichi target 11 and on comparisons of performance curves to efficiently represent taxa ranges delimited by SDM and proxies of genetic differentiation (Fig. 6). As using SDM combined with proxies of genetic differentiation showed the highest representation of genetic diversity (“SDM*PGD” scenario), we used this approach for the final analyses.Final conservation analysisWe identified areas of high conservation value for CWR in Mexico by using the software Zonation version 4.050,119, a systematic conservation planning tool that allows optimizing representation of species, taxa, or other conservation features, e.g., proxies of genetic differentiation, in a given study area. The program hierarchically ranks areas by removing cells of low conservation value, as given, for example, by a reduced number of taxa or occurrence of low weighted features, while considering multiple criteria such as the weighting of taxa and habitat preference of taxa. We applied the core-area zonation removal rule (CAZ) to maximize the representation of all conservation features in a minimal possible area51. Zonation generates two main outputs: (a) a hierarchical landscape priority rank map, that allows decision makers establishing different area thresholds to highlight areas of conservation interest; and (b) a representation curve showing species or conservation features range distribution in a given area. The curve also allows identifying how much area is needed to cover a certain taxon range or the distribution of a feature of conservation interest.For the conservation scenarios, we integrated the following inputs in the Zonation software: (1) 5,004 layers, i.e., SDM intersected with proxies of genetic differentiation (as described by “SDM*PGD” scenario, Fig. 4), (2) occurrence records of 98 taxa; only for those taxa without SDM, see Supplementary Data 3), (3) taxa specific habitat layers (according to Supplementary Data 11 and Supplementary Fig. 9), and (4) IUCN threat category (Supplementary Data 3) as an additional parameter to weight taxa differently to consider their vulnerability to extinction, see details below. See Zonation configuration at Supplementary Note 6.Data from different sources can be mixed in the same analysis, which is useful to not lose or omit information of any taxa of interest in the assessment. Here, we included information of a total of 214 taxa (see Supplementary Data 3). Distribution data of 116 taxa were represented by 5004 layers that resulted from combining 116 SDM and 102 PGD. This approach showed the highest proportion of area of taxa ranges (on average 41%) and highest representation of PGD within the area of each taxon (on average 76%; Fig. 4; see description in the main text). For some taxa, e.g. Cucurbita pepo, Physalis cinerascens, and Zea mays information on its distribution was assessed at subspecies level rather than at species level, explaining the difference in numbers of CWR taxa.In addition, we included occurrence data of 98 taxa without SDM to prevent missing important areas of taxa known distribution that are important to conserve (see Supplementary Data 3). We enabled the function ‘species of special interest’ (SSI) of Zonation, and included a SSI feature list file, listing the taxon names, as well as taxon-specific coordinate file for each of the 98 taxa that have been reviewed by the experts of each group. The spatial reference system was World Mercator projection. Occurrence data and SDM are treated similarly in the Zonation analysis, i.e., cells where taxa occur will be retained in the solution as long as possible to maximize its representation in the solution.We assigned weights to the 116 taxa with SDM by using IUCN threat categories (according to Supplementary Data 3), giving highest values to taxa with highest risk of extinction that urgently need management actions to further avoid genetic erosion. By including conservation feature weights, Zonation estimates the conservation value of a cell not only based on the presences of a taxa and their distribution range, but also on the weight. A high weight indicates a high conservation value of cells where these taxa are distributed. As there is no rule for weight setting, we assigned values between 1 and 0 regardless of taxa distribution ranges, which is automatically considered in the Zonation algorithm to guarantee the representation of locations where limited-range distributed taxa occur within the most valuable conservation area. Thus, weights were assigned as follows: Critically endangered, CR: 1; Endangered, EN: 1; Vulnerable, VU: 0.8; Near threatened, NT: 0.5; Data deficient, DD: 0.3; Least concern, LC: 0.2 Not evaluated, NE: 0.1. SSI taxa were all weighted similarly with 1 in order to represent the 98 SSI taxa and their occurrences in the top fraction of the most valuable conservation area, as these areas could be considered as ‘irreplaceable’ in terms of conservation. The conservation of these taxa that are only known in a few locations is crucial to maintain their populations. Information on weights for taxa with and without SDM is included in the file that lists the 5004 conservation features and the SSI file, respectively.To include the information on habitat, we included 116 habitat maps which guide the selection of cells to areas where its presence is more probable (see the Methods section: “Habitat preference”). This option can only be used for taxa represented by a raster layer, and is not available for SSI taxa included via occurrence records. By enabling the “landscape condition” option of Zonation, each habitat map is linked to a specific conservation feature layer. Areas with unfavorable habitats will quickly be masked out during the selection of cells in order to obtain a solution that favors conservation areas within areas of preferred habitat.We generated three final scenarios to identify conservation areas for (a) all taxa, (b) taxa exclusively distributing in natural vegetation, and (c) taxa associated with a wider range of habitats such as natural vegetation, agricultural and urban areas. The Zonation configuration remained similar among the three scenarios. When taxa were not included in a given scenario, we assigned a value weight of 0. This excluded the feature to be considered for the hierarchical prioritization of the landscape, but still allowed to evaluate the taxa during post-processing.To evaluate the spatial results (Supplementary Fig. 11), we analyzed performance curves to represent proxies of genetic differentiation within each taxon range (Supplementary Fig. 12). Also, we considered the most valuable 20% area of Mexico to calculate the coincidence of the three scenarios (Supplementary Fig. 13), and the overlap with federal protected areas125 and indigenous regions126,127 (Supplementary Fig. 14), and land cover data used in the analyses (Supplementary Figs. 9, 15).We discussed the proposed methodological framework, input layer and criteria during a fourth workshop in Mexico. It is worth mentioning that we ran several analyses including additional layers, such as areas where indigenous communities live that promote the presence of CWR in the landscape6. However, as the output indicated no evident difference by including this information, final analyses did not consider these data. We neither included protected areas nor tried to expand on the current 12% protected area system, because most management plans do not specifically address CWR management (but see the management program of the Protected Area of ‘Sierra de Manantlán’128), and thus generally do not adequately plan for wild and native genetic resources129. We also discussed different approaches to consider connectivity for taxa, habitats and proxies of genetic differentiation in the Zonation processing. Still, we finally decided to run the analysis without particularly accounting for connectivity as we had no taxa-specific information on dispersal abilities or possible effects of fragmentation, and we did not want to lose efficiency of the solution to represent taxa by or include lower-quality habitats by forcing the solution to an aggregation of pixels.Reporting summaryFurther information on research design is available in the Nature Research Reporting Summary linked to this article. More

Blatchford, M. L., Mannaerts, C. M., Zeng, Y., Nouri, H. & Karimi, P. Status of accuracy in remotely sensed and in-situ agricultural water productivity estimates: A review. Remote Sensing of Environment 234, 111413, https://doi.org/10.1016/j.rse.2019.111413 (2019).Article 
ADS 

Google Scholar 
Geerts, S. & Raes, D. Deficit irrigation as an on-farm strategy to maximize crop water productivity in dry areas. Agricultural Water Management 96, 1275–1284, https://doi.org/10.1016/j.agwat.2009.04.009 (2009).Article 

Google Scholar 
Hellegers, P., Soppe, R., Perry, C. & Bastiaanssen, W. Combining remote sensing and economic analysis to support decisions that affect water productivity. Irrigation Science 27, 243–251, https://doi.org/10.1007/s00271-008-0139-7 (2009).Article 

Google Scholar 
Bastiaanssen, W. G. M. & Steduto, P. The water productivity score (WPS) at global and regional level: Methodology and first results from remote sensing measurements of wheat, rice and maize. The Science of the total environment 575, https://doi.org/10.1016/j.scitotenv.2016.09.032 (2017).Seneviratne, S. I. et al. Investigating soil moisture–climate interactions in a changing climate: A review. Earth Science Reviews 99, https://doi.org/10.1016/j.earscirev.2010.02.004 (2010).Hu, X., Shi, L., Lin, L. & Zha, Y. Nonlinear boundaries of land surface temperature–vegetation index space to estimate water deficit index and evaporation fraction. Agricultural and Forest Meteorology 279, https://doi.org/10.1016/j.agrformet.2019.107736 (2019).Bowen, I. S. The Ratio of Heat Losses by Conduction and by Evaporation from any Water Surface. Physical Review 27, 779–787, https://doi.org/10.1103/PhysRev.27.779 (1926).Article 
ADS 
CAS 
MATH 

Google Scholar 
Penman, H. L. Natural evaporation from open water, hare soil and grass. Proceedings of the Royal Society of London. Series A, Mathematical and physical sciences 193, https://doi.org/10.1098/rspa.1948.0037 (1948).Monteith, J. L. Evaporation and environment. The stage and movement of water in living organisms. Symp.soc.exp.biol.the Company of Biologists (1965).Wang, K. & Dickinson, R. E. A review of global terrestrial evapotranspiration: Observation, modeling, climatology, and climatic variability. Reviews of Geophysics 50, https://doi.org/10.1029/2011RG000373 (2012).Bastiaanssen, W. G. et al. A remote sensing surface energy balance algorithm for land (SEBAL) Part 1: Fomulation. Journal of hydrology 212, 213–229, https://doi.org/10.1016/S0022-1694(98)00253-4 (1998).Article 
ADS 

Google Scholar 
Bastiaanssen, W. G. M. et al. A remote sensing surface energy balance algorithm for land (SEBAL) Part 2. Validation. Journal of Hydrology 212, https://doi.org/10.1016/S0022-1694(98)00254-6 (1998).Su, Z. The Surface Energy Balance System (SEBS) for estimation of turbulent heat fluxes. Hydrology and Earth System Science 6, 85–99, https://doi.org/10.5194/hess-6-85-2002 (2002).Article 
ADS 

Google Scholar 
Norman, J. M., Kustas, W. P. & Humes, K. S. Source approach for estimating soil and vegetation energy fluxes in observations of directional radiometric surface temperature. Agricultural and Forest Meteorology 77, https://doi.org/10.1016/0168-1923(95)02265-y (1995).Mu, Q., Heinsch, F. A., Zhao, M. & Running, S. W. Development of a global evapotranspiration algorithm based on MODIS and global meteorology data. Remote Sensing of Environment 111, https://doi.org/10.1016/j.rse.2007.04.015 (2007).Mu, Q., Zhao, M. & Running, S. W. Improvements to a MODIS global terrestrial evapotranspiration algorithm. Remote Sensing of Environment 115, 1781–1800, https://doi.org/10.1016/j.rse.2011.02.019 (2011).Article 
ADS 

Google Scholar 
Fisher, J. B., Tu, K. P. & Baldocchi, D. D. Global estimates of the land–atmosphere water flux based on monthly AVHRR and ISLSCP-II data, validated at 16 FLUXNET sites. Remote Sensing of Environment 112, 901–919, https://doi.org/10.1016/j.rse.2007.06.025 (2008).Article 
ADS 

Google Scholar 
Kim, H. W., Hwang, K., Mu, Q., Lee, S. O. & Choi, M. Validation of MODIS 16 global terrestrial evapotranspiration products in various climates and land cover types in Asia. KSCE Journal of Civil Engineering 16, https://doi.org/10.1007/s12205-012-0006-1 (2012).Velpuri, N. M., Senay, G. B., Singh, R. K., Bohms, S. & Verdin, J. P. A comprehensive evaluation of two MODIS evapotranspiration products over the conterminous United States: Using point and gridded FLUXNET and water balance ET. Remote Sensing of Environment 139, https://doi.org/10.1016/j.rse.2013.07.013 (2013).Jin, X. et al. Estimation of water productivity in winter wheat using the AquaCrop model with field hyperspectral data. Precision Agriculture 19, 1–17, https://doi.org/10.1007/s11119-016-9469-2 (2016).Article 

Google Scholar 
Felix, R., Clement, A., Igor, S. & Oscar, R. Using Low Resolution Satellite Imagery for Yield Prediction and Yield Anomaly Detection. Remote Sensing 5, 1704–1733, https://doi.org/10.3390/rs5041704 (2013).Article 

Google Scholar 
Lu, Y. et al. Assimilation of soil moisture and canopy cover data improves maize simulation using an under-calibrated crop model. Agricultural Water Management 252, https://doi.org/10.1016/j.agwat.2021.106884 (2021).Jin, X., Kumar, L., Li, Z., Feng, H. & Wang, J. A review of data assimilation of remote sensing and crop models. European Journal of Agronomy 92, https://doi.org/10.1016/j.eja.2017.11.002 (2018).Weiss, M., Jacob, F. & Duveiller, G. Remote sensing for agricultural applications: A meta-review. Remote Sensing of Environment 236, https://doi.org/10.1016/j.rse.2019.111402 (2019).Jin, X. et al. Winter wheat yield estimation based on multi-source medium resolution optical and radar imaging data and the AquaCrop model using the particle swarm optimization algorithm. ISPRS Journal of Photogrammetry and Remote Sensing 126, 24–37 (2017).Article 
ADS 

Google Scholar 
Tao, F., Rötter, R. P., Palosuo, T., Díaz-Ambrona, C. G. H. & Schulman, A. H. Contribution of crop model structure, parameters and climate projections to uncertainty in climate change impact assessments. Global Change Biology 24, https://doi.org/10.1111/gcb.14019 (2017).Jin, X. et al. A review of data assimilation of remote sensing and crop models. European Journal of Agronomy 92, 141–152, https://doi.org/10.1016/j.eja.2017.11.002 (2018).Article 

Google Scholar 
Anikó, K. et al. Statistical modelling of crop yield in Central Europe using climate data and remote sensing vegetation indices. Agricultural and Forest Meteorology 260-261, 300–320, https://doi.org/10.1016/j.agrformet.2018.06.009 (2018).Article 

Google Scholar 
Wang, Y., Zhang, Z., Feng, L., Du, Q. & Runge, T. Combining Multi-Source Data and Machine Learning Approaches to Predict Winter Wheat Yield in the Conterminous United States. Remote Sensing 12, 1232, https://doi.org/10.3390/rs12081232 (2020).Article 
ADS 

Google Scholar 
Franz, T. E. et al. The role of topography, soil, and remotely sensed vegetation condition towards predicting crop yield. Field Crops Research 252, https://doi.org/10.1016/j.fcr.2020.107788 (2020).Noland, R. L. et al. Estimating alfalfa yield and nutritive value using remote sensing and air temperature. Field Crops Research 222, 189–196, https://doi.org/10.1016/j.fcr.2018.01.017 (2018).Article 

Google Scholar 
Cao, J., Zhang, Z., Luo, Y., Zhang, L. & Tao, F. Wheat yield predictions at a county and field scale with deep learning, machine learning, and google earth engine. European Journal of Agronomy, 126204, https://doi.org/10.1016/j.eja.2020.126204 (2021).Jacinta, H. & Kerrie, M. Statistical Machine Learning Methods and Remote Sensing for Sustainable Development Goals: A Review. Remote Sensing 10, 1365, https://doi.org/10.3390/rs10091365 (2018).Article 

Google Scholar 
Jin, X., Liu, S., Baret, F., Hemerlé, M. & Comar, A. Estimates of plant density of wheat crops at emergence from very low altitude UAV imagery. Remote Sensing of Environment 198, 105–114, https://doi.org/10.1016/j.rse.2017.06.007 (2017).Article 
ADS 

Google Scholar 
Maimaitijiang, M. et al. Soybean yield prediction from UAV using multimodal data fusion and deep learning. Remote Sensing of Environment 237, 111599, https://doi.org/10.1016/j.rse.2019.111599 (2020).Article 
ADS 

Google Scholar 
Hossein, A., Mohsen, A., Davoud, A., Salehi, S. H. & Soheil, R. Machine Learning Regression Techniques for the Silage Maize Yield Prediction Using Time-Series Images of Landsat 8 OLI. IEEE Journal of Selected Topics in Applied Earth Observations Remote Sensing PP, 1–15, https://doi.org/10.1109/JSTARS.2018.2823361 (2018).Johansen, K. et al. Predicting Biomass and Yield in a Tomato Phenotyping Experiment Using UAV Imagery and Random Forest. Frontiers in Artificial Intelligence 3, 28, https://doi.org/10.3389/frai.2020.00028 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Zhang, L., Ding, X., Shen, Y., Wang, Z. & Wang, X. Spatial Heterogeneity and Influencing Factors of Agricultural Water Use Efficiency in China. Resources and Environment in the Yangtze Basin 28, https://doi.org/10.11870/cjlyzyyhj201904008 (2019).Cheng, M. et al. Satellite time series data reveal interannual and seasonal spatiotemporal evapotranspiration patterns in China in response to effect factors. Agric. Water Manage. 255, https://doi.org/10.1016/j.agwat.2021.107046 (2021).Zhou, L. Comprehensive agricultural regionalization in China. (Agricultural Press of China, 1985).Luo, Y., Zhang, Z., Chen, Y., Li, Z. & Tao, F. ChinaCropPhen1km: A high-resolution crop phenological dataset for three staple crops in China during 2000-2015 based on LAI products. Figshare https://doi.org/10.6084/m9.figshare.8313530.v6 (2019).Luo, Y., Zhang, Z., Chen, Y., Li, Z. & Tao, F. ChinaCropPhen1km: a high-resolution crop phenological dataset for three staple crops in China during 2000–2015 based on leaf area index (LAI) products. Earth System Science Data 12, 197–214, https://doi.org/10.5194/essd-12-197-2020 (2020).Article 
ADS 

Google Scholar 
Song, D. Second China Soil Survey. (Chinese Science Press, 1979).Zhang, T., Yang, X., Wang, H., Li, Y. & Ye, Q. Climatic and technological ceilings for Chinese rice stagnation based on yield gaps and yield trend pattern analysis. Global Change Biology 20, 1289–1298, https://doi.org/10.1111/gcb.12428 (2014).Article 
ADS 
PubMed 

Google Scholar 
Chen, Y., Zhang, Z. & Tao, F. Improving regional winter wheat yield estimation through assimilation of phenology and leaf area index from remote sensing data. European Journal of Agronomy 101, 163–173, https://doi.org/10.1016/j.eja.2018.09.006 (2018).Article 

Google Scholar 
Cheng, M. et al. Combining multi-indicators with machine-learning algorithms for maize yield early prediction at the county-level in China. Agricultural and Forest Meteorology 323, https://doi.org/10.1016/j.agrformet.2022.109057 (2022).Amir, J. & Sinclair, T. A model of the temperature and solar-radiation effects on spring wheat growth and yield. Field Crops Research 28, 47–58, https://doi.org/10.1016/0378-4290(91)90073-5 (1991).Article 

Google Scholar 
Prince, S. D., Haskett, J., Steininger, M. & Wright, S. R. Net Primary Production of U.S. Midwest Croplands from Agricultural Harvest Yield Data. Ecological Applications 11, 1194–1205, https://doi.org/10.1890/1051-0761(2001)011[1194:NPPOUS]2.0.CO;2 (2001).Article 

Google Scholar 
Gilardelli, C. et al. Downscaling rice yield simulation at sub-field scale using remotely sensed LAI data. European journal of agronomy 103, 108–116, https://doi.org/10.1016/j.eja.2018.12.003 (2019).Article 

Google Scholar 
Shakoor, R., Hassan, M. Y., Raheem, A. & Wu, Y.-K. Wake effect modeling: A review of wind farm layout optimization using Jensen׳ s model. Renewable and Sustainable Energy Reviews 58, 1048–1059, https://doi.org/10.1016/j.rser.2015.12.229 (2016).Article 

Google Scholar 
Breiman, L. Random Forests. Machine Learning https://doi.org/10.1023/A:1010933404324 (2001).Article 
MATH 

Google Scholar 
Li, L. et al. Crop yield forecasting and associated optimum lead time analysis based on multi-source environmental data across China. Agricultural and Forest Meteorology 308–309, https://doi.org/10.1016/j.agrformet.2021.108558 (2021).Wang, L. A., Zhou, X., Zhu, X., Dong, Z. & Guo, W. Estimation of biomass in wheat using random forest regression algorithm and remote sensing data. The Crop Journal 4, 212–219, https://doi.org/10.1016/j.cj.2016.01.008 (2016).Article 

Google Scholar 
Feng, P. et al. Dynamic wheat yield forecasts are improved by a hybrid approach using a biophysical model and machine learning technique. Agricultural and Forest Meteorology 285-286, 107922, https://doi.org/10.1016/j.agrformet.2020.107922 (2020).Article 
ADS 

Google Scholar 
Lu, F., Sun, Y. & Hou, F. Using UAV Visible Images to Estimate the Soil Moisture of Steppe. Water 12, 2334, https://doi.org/10.3390/w12092334 (2020).Article 
CAS 

Google Scholar 
Wang, S. et al. High spatial resolution monitoring land surface energy, water and CO2 fluxes from an Unmanned Aerial System. Remote Sensing of Environment 229, 14–31, https://doi.org/10.1016/j.rse.2019.03.040 (2019).Article 
ADS 

Google Scholar 
Chen, Y. et al. Comparison of satellite-based evapotranspiration models over terrestrial ecosystems in China. Remote Sensing of Environment 140, 279–293, https://doi.org/10.1016/j.rse.2013.08.045 (2014).Article 
ADS 

Google Scholar 
Peralta, N., Assefa, Y., Du, J., Barden, C. & Ciampitti, I. Mid-Season High-Resolution Satellite Imagery for Forecasting Site-Specific Corn Yield. Remote Sensing 8, 848, https://doi.org/10.3390/rs8100848 (2016).Article 
ADS 

Google Scholar 
Russello, H. Convolutional neural networks for crop yield prediction using satellite images. IBM Center for Advanced Studies (2018).You, J., Li, X., Low, M., Lobell, D. & Ermon, S. in Proceedings of the AAAI Conference on Artificial Intelligence.Moran, P. A. Notes on continuous stochastic phenomena. Biometrika 37, 17–23 (1950).Article 
MathSciNet 
CAS 

Google Scholar 
Imran, M., Stein, A. & Zurita-Milla, R. Using geographically weighted regression kriging for crop yield mapping in West Africa. International Journal of Geographical Information Systems 29, 234–257, https://doi.org/10.1080/13658816.2014.959522 (2015).Article 

Google Scholar 
Harries, K. Extreme spatial variations in crime density in Baltimore County, MD. Geoforum 37, 404–416, https://doi.org/10.1016/j.geoforum.2005.09.004 (2006).Article 

Google Scholar 
Ghulam, A. et al. Remote Sensing Based Spatial Statistics to Document Tropical Rainforest Transition Pathways. Remote Sensing 7, 6257–6279, https://doi.org/10.3390/rs70506257 (2015).Article 
ADS 

Google Scholar 
Maimaitijiang, M., Ghulam, A., Sandoval, J. S. O. & Maimaitiyiming, M. Drivers of land cover and land use changes in St. Louis metropolitan area over the past 40 years characterized by remote sensing and census population data. International Journal of Applied Earth Observation Geoinformation 35, 161–174, https://doi.org/10.1016/j.jag.2014.08.020 (2015).Article 
ADS 

Google Scholar 
Cheng, M. Long time series (2001-2015) high-resolution crop yield and water productivity dataset of China, Zenodo, https://doi.org/10.5281/zenodo.5121842 (2021).Martens, B., Miralles, D. G., Lievens, H., Schalie, R. D. & Verhoest, N. GLEAM v3: Satellite-based land evaporation and root-zone soil moisture. Geoscientific Model Development 10, https://doi.org/10.5194/gmd-10-1903-2017 (2017).Wang, W., Cui, W., Wang, X. & Chen, X. Evaluation of GLDAS-1 and GLDAS-2 forcing data and Noah model simulations over China at the monthly scale. Journal of Hydrometeorology 17, 2815–2833, https://doi.org/10.1175/JHM-D-15-0191.1 (2016).Article 
ADS 

Google Scholar 
Chen, X. et al. Development of a 10-year (2001–2010) 0.1° data set of land-surface energy balance for mainland China. Atmospheric Chemistry and Physics 14, 14471–14518, https://doi.org/10.5194/acp-14-13097-2014 (2014).Article 
ADS 
CAS 

Google Scholar 
Ramoelo, A. et al. Validation of Global Evapotranspiration Product (MOD16) using Flux Tower Data in the African Savanna, South Africa. Remote Sensing 6, https://doi.org/10.3390/rs6087406 (2014).Yang, X., Yong, B., Ren, L., Zhang, Y. & Long, D. Multi-scale validation of GLEAM evapotranspiration products over China via ChinaFLUX ET measurements. International Journal of Remote Sensing https://doi.org/10.1080/01431161.2017.1346400 (2017).Article 

Google Scholar 
Hu, G., Jia, L. & Menenti, M. Comparison of MOD16 and LSA-SAF MSG evapotranspiration products over Europe for 2011. Remote Sensing of Environment 156, 510–526, https://doi.org/10.1016/j.rse.2014.10.017 (2015).Article 
ADS 

Google Scholar 
Khan, M. S., Liaqat, U. W., Baik, J. & Choi, M. Stand-alone uncertainty characterization of GLEAM, GLDAS and MOD16 evapotranspiration products using an extended triple collocation approach. Agricultural and Forest Meteorology 252, 256–268, https://doi.org/10.1016/j.agrformet.2018.01.022 (2018).Article 
ADS 

Google Scholar 
Glenn, E. P. et al. Scaling sap flux measurements of grazed and ungrazed shrub communities with fine and coarse-resolution remote sensing. Ecohydrology 1, 316–329, https://doi.org/10.1002/eco.19 (2008).Article 

Google Scholar 
Gamon, J. A. Reviews and Syntheses: optical sampling of the flux tower footprint. Biogeosciences 12, 4509–4523, https://doi.org/10.5194/bg-12-4509-2015 (2015).Article 
ADS 

Google Scholar 
Cai, Y. et al. Integrating satellite and climate data to predict wheat yield in Australia using machine learning approaches. Agricultural and Forest Meteorology 274, 144–159, https://doi.org/10.1016/j.agrformet.2019.03.010 (2019).Article 
ADS 

Google Scholar 
Chen, X. et al. Prediction of Maize Yield at the City Level in China Using Multi-Source Data. Remote Sensing 13, https://doi.org/10.3390/rs13010146 (2021).Guo, Y. et al. Integrated phenology and climate in rice yields prediction using machine learning methods. Ecological Indicators 120, 106935, https://doi.org/10.1016/j.ecolind.2020.106935 (2021).Article 

Google Scholar 
Yuan, W. et al. Estimating crop yield using a satellite-based light use efficiency model. Ecological Indicators 60, 702–709, https://doi.org/10.1016/j.ecolind.2015.08.013 (2016).Article 

Google Scholar 
Anandhi, A. Growing degree days – Ecosystem indicator for changing diurnal temperatures and their impact on corn growth stages in Kansas. Ecological Indicators 61, 149–158, https://doi.org/10.1016/j.ecolind.2015.08.023 (2016).Article 

Google Scholar 
Wart, J. V. Estimating Crop Yield Potential At National Scales. Field Crops Research 143, 34–43, https://doi.org/10.1016/j.fcr.2012.11.018 (2013).Article 

Google Scholar 
Kang, Y. S. et al. Yield prediction and validation of onion (Allium cepa L.) using key variables in narrowband hyperspectral imagery and effective accumulated temperature. Computers and Electronics in Agriculture 178, https://doi.org/10.1016/j.compag.2020.105667 (2020).Long, D., Singh, V. P. & Li, Z.-L. How sensitive is SEBAL to changes in input variables, domain size and satellite sensor? Journal of Geophysical Research: Atmospheres 116, https://doi.org/10.1029/2011jd016542 (2011).Liu, Z., Wang, L. & Wang, S. Comparison of Different GPP Models in China Using MODIS Image and ChinaFLUX Data. Remote Sensing 6, 10215–10231, https://doi.org/10.3390/rs61010215 (2014).Article 
ADS 

Google Scholar 
Edreira, J., Guilpart, N., Sadras, V., Cassman, K. G. & Grassini, P. Water productivity of rainfed maize and wheat: A local to global perspective. Agricultural and Forest Meteorology 259, 364–373, https://doi.org/10.1016/j.agrformet.2018.05.019 (2018).Article 
ADS 

Google Scholar 
Li, H. et al. Water Use Characteristics of Maize-Green Manure Intercropping Under Different Nitrogen Application Levels in the Oasis Irrigation Area Scientia Agricultura Sinica 54, 2608–2618 (2021).
Google Scholar 
Wang, S., Ibrom, A., Bauer-Gottwein, P. & Garcia, M. Incorporating diffuse radiation into a light use efficiency and evapotranspiration model: An 11-year study in a high latitude deciduous forest. Agricultural and Forest Meteorology https://doi.org/10.1016/j.agrformet.2017.10.023 (2018).Article 

Google Scholar 
Cheng, M. High-resolution crop yield and water productivity dataset generated using random forest and remote sensing. Zenodo https://doi.org/10.5281/zenodo.6444614 (2022). More

Maxwell, S. L., Fuller, R. A., Brooks, T. M. & Watson, J. E. M. Biodiversity: The ravages of guns, nets and bulldozers. Nature 536, 143–145 (2016).ADS 
CAS 
PubMed 

Google Scholar 
Uchida, K. & Ushimaru, A. Biodiversity declines due to abandonment and intensification of agricultural lands: Patterns and mechanisms. Ecol. Monogr. 84, 637–658 (2014).
Google Scholar 
Habel, J. C. et al. Butterfly community shifts over two centuries: Shifts in butterfly communities. Conserv. Biol. 30, 754–762 (2016).PubMed 

Google Scholar 
Seibold, S. et al. Arthropod decline in grasslands and forests is associated with landscape-level drivers. Nature 574, 671–674 (2019).ADS 
CAS 
PubMed 

Google Scholar 
Hallmann, C. A. et al. More than 75 percent decline over 27 years in total flying insect biomass in protected areas. PLoS One 12, e0185809 (2017).PubMed 
PubMed Central 

Google Scholar 
Wenzel, M., Schmitt, T., Weitzel, M. & Seitz, A. The severe decline of butterflies on western German calcareous grasslands during the last 30 years: A conservation problem. Biol. Cons. 128, 542–552 (2006).
Google Scholar 
Biesmeijer, J. C. et al. Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313, 351–354 (2006).ADS 
CAS 
PubMed 

Google Scholar 
Hallmann, C. A., Foppen, R. P. B., van Turnhout, C. A. M., de Kroon, H. & Jongejans, E. Declines in insectivorous birds are associated with high neonicotinoid concentrations. Nature 511, 341–343 (2014).ADS 
CAS 
PubMed 

Google Scholar 
Møller, A. P. Parallel declines in abundance of insects and insectivorous birds in Denmark over 22 years. Ecol. Evol. 9, 6581–6587 (2019).PubMed 
PubMed Central 

Google Scholar 
Wagner, D. L. Insect declines in the anthropocene. Annu. Rev. Entomol. 65, 457–480 (2020).CAS 
PubMed 

Google Scholar 
Habel, J. C., Samways, M. J. & Schmitt, T. Mitigating the precipitous decline of terrestrial European insects: Requirements for a new strategy. Biodivers. Conserv. 28, 1343–1360 (2019).
Google Scholar 
Uhl, B., Wölfling, M. & Fiedler, K. Understanding small-scale insect diversity patterns inside two nature reserves: The role of local and landscape factors. Biodivers. Conserv. 29, 2399–2418 (2020).
Google Scholar 
Stevens, C. J., Dise, N. B., Mountford, J. O. & Gowing, D. J. Impact of nitrogen deposition on the species richness of grasslands. Science 303, 1876–1879 (2004).ADS 
CAS 
PubMed 

Google Scholar 
Thomas, J. A. Butterfly communities under threat. Science 353, 216–218 (2016).ADS 
CAS 
PubMed 

Google Scholar 
Sanders, J. & Hess, J. Benefits of organic farming to environment and society. Thünen Report 65, 362 (2019).
Google Scholar 
Brühl, C. A. & Zaller, J. G. Biodiversity decline as a consequence of an inappropriate environmental risk assessment of pesticides. Front. Environ. Sci. 7, 177 (2019).
Google Scholar 
Brühl, C. A. et al. Direct pesticide exposure of insects in nature conservation areas in Germany. Sci. Rep. 11, 24144 (2021).ADS 
PubMed 
PubMed Central 

Google Scholar 
Wagner, D. L., Grames, E. M., Forister, M. L., Berenbaum, M. R. & Stopak, D. Insect decline in the Anthropocene: Death by a thousand cuts. Proc. Natl. Acad. Sci. USA 118, e2023989118 (2021).CAS 
PubMed 
PubMed Central 

Google Scholar 
Den Boer, P. J. & van Dijk, T. S. Carabid Beetles in A Changing Environment (Agricultural Univ, 1995).
Google Scholar 
Cristescu, M. E. From barcoding single individuals to metabarcoding biological communities: Towards an integrative approach to the study of global biodiversity. Trends Ecol. Evol. 29, 566–571 (2014).PubMed 

Google Scholar 
Hausmann, A. et al. Toward a standardized quantitative and qualitative insect monitoring scheme. Ecol. Evol. 10, 4009–4020 (2020).PubMed 
PubMed Central 

Google Scholar 
Ratnasingham, S. & Hebert, P. D. N. A DNA-based registry for all animal species: The Barcode Index Number (BIN) system. PLoS One 8, e66213 (2013).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Hausmann, A. et al. Genetic patterns in european geometrid moths revealed by the Barcode Index Number (BIN) system. PLoS One 8, e84518 (2013).ADS 
PubMed 
PubMed Central 

Google Scholar 
Padial, J. M., Miralles, A., De la Riva, I. & Vences, M. The integrative future of taxonomy. Front. Zool. 7, 1–14 (2010).
Google Scholar 
Schlick-Steiner, B. C. et al. Integrative taxonomy: A multisource approach to exploring biodiversity. Ann. Rev. Entomol. 55, 421–438 (2010).CAS 

Google Scholar 
Schlick‐Steiner, B. C., Arthofer, W., & Steiner, F. M. Take up the challenge! Opportunities for evolution research from resolving conflict in integrative taxonomy (2014).Fujita, M. K., Leaché, A. D., Burbrink, F. T., McGuire, J. A. & Moritz, C. Coalescent-based species delimitation in an integrative taxonomy. Trends Ecol. Evol. 27, 480–488 (2012).PubMed 

Google Scholar 
Morinière, J. et al. A DNA barcode library for 5,200 German flies and midges (Insecta: Diptera) and its implications for metabarcoding-based biomonitoring. Mol. Ecol. Res. 19, 900–928 (2019).
Google Scholar 
Kortmann, M. et al. Arthropod dark taxa provide new insights into diversity responses to bark beetle infestations. Ecol. Appl. 32, e2516 (2022).PubMed 

Google Scholar 
Porter, T. M. & Hajibabaei, M. Automated high throughput animal CO1 metabarcode classification. Sci. Rep. 8, 1–10 (2018).
Google Scholar 
Boggs, C. L. & Inouye, D. W. A single climate driver has direct and indirect effects on insect population dynamics: Climate drivers of population dynamics. Ecol. Lett. 15, 502–508 (2012).PubMed 

Google Scholar 
Conrad, K. F., Fox, R. & Woiwod, I. P. Monitoring biodiversity: Measuring long-term changes in insect abundance. In Insect Conservation Biology (eds Stewart, A. J. A. et al.) 203–225 (CABI, 2007). https://doi.org/10.1079/9781845932541.0203.Chapter 

Google Scholar 
Flohre, A. et al. Agricultural intensification and biodiversity partitioning in European landscapes comparing plants, carabids, and birds. Ecol. Appl. Publ. Ecol. Soc. Am. 21, 1772–1781 (2011).
Google Scholar 
Emmerson, M. et al. How agricultural intensification affects biodiversity and ecosystem services. In Advances in Ecological Research, vol ***55 43–97 (Elsevier, 2016).
Google Scholar 
Segerer, A. H. & Rosenkranz, E. Das grosse Insektensterben: Was es Bedeutet und was Wir Jetzt tun Müssen (Oekom Verlag, 2019).
Google Scholar 
Batáry, et al. The former Iron Curtain still drives biodiversity-profit trade-offs in German agriculture. Nat. Ecol. Evol. 1, 1279–1284 (2017).PubMed 

Google Scholar 
Kuussaari, M. et al. Extinction debt: A challenge for biodiversity conservation. Trends Ecol. Evol. 24, 564–571 (2009).PubMed 

Google Scholar 
Birkhofer, K., Smith, H. G., Weisser, W. W., Wolters, V. & Gossner, M. M. Land-use effects on the functional distinctness of arthropod communities. Ecography 38, 889–900 (2015).
Google Scholar 
Tscharntke, T., Klein, A. M., Kruess, A., Steffan-Dewenter, I. & Thies, C. Landscape perspectives on agricultural intensification and biodiversity—ecosystem service management. Ecol. Lett. 8, 857–874 (2005).
Google Scholar 
Habel, J. C., Seibold, S., Ulrich, W. & Schmitt, T. Seasonality overrides differences in butterfly species composition between natural and anthropogenic forest habitats. Anim. Conserv. 21, 405–413 (2018).
Google Scholar 
Schmitt, T., Ulrich, W., Delic, A., Teucher, M. & Habel, J. C. Seasonality and landscape characteristics impact species community structure and temporal dynamics of East African butterflies. Sci. Rep. 11, 15103 (2021).CAS 
PubMed 
PubMed Central 

Google Scholar 
Ssymank, A. et al. Praktische Hinweise und Empfehlungen zur Anwendung von Malaisefallen für Insekten in der Biodiversitätserfassung und im Monitoring. Entomol. Verein Krefeld 1, 1–12 (2018).
Google Scholar 
Elbrecht, V., Peinert, B. & Leese, F. Sorting things out: Assessing effects of unequal specimen biomass on DNA metabarcoding. Ecol. Evol. 7, 6918–6926 (2017).PubMed 
PubMed Central 

Google Scholar 
Elbrecht, V. & Steinke, D. Scaling up DNA metabarcoding for freshwater macrozoobenthos monitoring. Freshw. Biol. 64, 380–387 (2019).CAS 

Google Scholar 
Boetzl, F. A. et al. A multitaxa assessment of the effectiveness of agri-environmental schemes for biodiversity management. Proc. Natl. Acad. Sci. 118, 25 (2021).
Google Scholar 
Uhler, J. et al. Relationship of insect biomass and richness with land use along a climate gradient. Nat. Commun. 12, 1–9 (2021).
Google Scholar 
Leray, M. et al. A new versatile primer set targeting a short fragment of the mitochondrial COI region for metabarcoding metazoan diversity: Application for characterizing coral reef fish gut contents. Front. Zool. 10, 34 (2013).PubMed 
PubMed Central 

Google Scholar 
Morinière, J. et al. Species identification in malaise trap samples by DNA barcoding based on NGS Technologies and a scoring matrix. PLoS One 11, e0155497 (2016).PubMed 
PubMed Central 

Google Scholar 
Rognes, T., Flouri, T., Nichols, B., Quince, C. & Mahé, F. VSEARCH: A versatile open source tool for metagenomics. PeerJ 4, e2584 (2016).PubMed 
PubMed Central 

Google Scholar 
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 17, 10 (2011).
Google Scholar 
Ondov, B. D., Bergman, N. H. & Phillippy, A. M. Interactive metagenomic visualization in a Web browser. BMC Bioinform. 12, 385 (2011).
Google Scholar  More

Goodale, E., Beauchamp, G. & Ruxton, G. D. Mixed-Species Groups of Animals: Behavior, Community Structure, and Conservation (Academic Press, 2017).
Google Scholar 
Krause, J. & Ruxton, G. D. Living in Groups (Oxford University Press, 2002).
Google Scholar 
Stensland, E., Angerbjorn, A. & Berggren, P. Mixed species groups in mammals. Mamm. Rev. 33, 205–223 (2003).
Google Scholar 
Anderson, T. M. et al. Landscape-scale analyses suggest both nutrient and antipredator advantages to Serengeti herbivore hotspots. Ecology 91, 1519–1529 (2010).PubMed 

Google Scholar 
Sinclair, A. R. E. Does interspecific competition or predation shape the African ungulate community? J. Anim. Ecol. 54, 899–918 (1985).
Google Scholar 
Kiffner, C., Kioko, J., Leweri, C. & Krause, S. Seasonal patterns of mixed species groups in large East African mammals. PLoS ONE 9, e113446 (2014).ADS 
PubMed 
PubMed Central 

Google Scholar 
Meise, K., Franks, D. W. & Bro-Jørgensen, J. Using social network analysis of mixed species groups in African savannah herbivores to assess how community structure responds to environmental change. Philos. Trans. R. Soc. B Biol. Sci. 374, 20190009 (2019).
Google Scholar 
de Boer, W. F. & Prins, H. H. T. Large herbivores that thrive mightily but eat and drink as friends. Oecologia 82, 264–274 (1990).ADS 
PubMed 

Google Scholar 
Beaudrot, L., Palmer, M. S., Anderson, T. M. & Packer, C. Mixed-species groups of Serengeti grazers: A test of the stress gradient hypothesis. Ecology. https://doi.org/10.1002/ecy.3163 (2020).Article 
PubMed 

Google Scholar 
He, Q., Bertness, M. D. & Altieri, A. H. Global shifts towards positive species interactions with increasing environmental stress. Ecol. Lett. 16, 695–706 (2013).PubMed 

Google Scholar 
Bertness, M. D. & Callaway, R. Positive interactions in communities. Trends Ecol. Evol. 9, 191–193 (1994).CAS 
PubMed 

Google Scholar 
Fugère, V. et al. Testing the stress-gradient hypothesis with aquatic detritivorous invertebrates: Insights for biodiversity-ecosystem functioning research. J. Anim. Ecol. 81, 1259–1267 (2012).PubMed 

Google Scholar 
Bakker, E. S., Dobrescu, I., Straile, D. & Holmgren, M. Testing the stress gradient hypothesis in herbivore communities: Facilitation peaks at intermediate nutrient levels. Ecology 94, 1776–1784 (2013).PubMed 

Google Scholar 
Hopcraft, J. G. C., Olff, H. & Sinclair, A. R. E. Herbivores, resources and risks: Alternating regulation along primary environmental gradients in savannas. Trends Ecol. Evol. 25, 119–128 (2010).PubMed 

Google Scholar 
Sih, A. Optimal behavior: Can foragers balance two conflicting demands? Science 210, 1041–1043 (1980).ADS 
CAS 
PubMed 

Google Scholar 
Creel, S. & Christianson, D. Relationships between direct predation and risk effects. Trends Ecol. Evol. 23, 194–201 (2008).PubMed 

Google Scholar 
Zollner, P. A. & Lima, S. L. Towards a behavioral ecology of ecological landscapes. Trends Ecol. Evol. 11, 131–135 (1996).PubMed 

Google Scholar 
Brown, J. S., Laundré, J. W. & Gurung, M. The ecology of fear: Optimal foraging, game theory, and trophic interactions. J. Mammal. 80, 385–399 (1999).
Google Scholar 
Gaynor, K. M., Brown, J. S., Middleton, A. D., Power, M. E. & Brashares, J. S. Landscapes of fear: Spatial patterns of risk perception and response. Trends Ecol. Evol. 34, 355–368 (2019).PubMed 

Google Scholar 
Creel, S., Schuette, P. & Christianson, D. Effects of predation risk on group size, vigilance, and foraging behavior in an African ungulate community. Behav. Ecol. 25, 773–784 (2014).
Google Scholar 
Goodale, E., Beauchamp, G., Magrath, R. D., Nieh, J. C. & Ruxton, G. D. Interspecific information transfer influences animal community structure. Trends Ecol. Evol. 25, 354–361 (2010).PubMed 

Google Scholar 
Freeberg, T. M., Eppert, S. K., Sieving, K. E. & Lucas, J. R. Diversity in mixed species groups improves success in a novel feeder test in a wild songbird community. Sci. Rep. 7, 43014 (2017).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Anderson, T. M. et al. The spatial distribution of african savannah herbivores: Species associations and habitat occupancy in a landscape context. Philos. Trans. R. Soc. B Biol. Sci. 371, 20150314 (2016).
Google Scholar 
Arsenault, R. & Owen-Smith, N. Resource partitioning by grass height among grazing ungulates does not follow body size relation. Oikos 117, 1711–1717 (2008).
Google Scholar 
Esmaeili, S. et al. Body size and digestive system shape resource selection by ungulates: A cross-taxa test of the forage maturation hypothesis. Ecol. Lett. 24, 2178–2191 (2021).PubMed 

Google Scholar 
Hopcraft, J. G. C., Anderson, T. M., Pérez-Vila, S., Mayemba, E. & Olff, H. Body size and the division of niche space: Food and predation differentially shape the distribution of Serengeti grazers. J. Anim. Ecol. 81, 201–213 (2012).PubMed 

Google Scholar 
McArthur, C., Banks, P. B., Boonstra, R. & Forbey, J. S. The dilemma of foraging herbivores: Dealing with food and fear. Oecologia 176, 677–689 (2014).ADS 
PubMed 

Google Scholar 
Gagnon, M. & Chew, A. E. Dietary preferences in extant African Bovidae. J. Mammal. 81, 490–511 (2000).
Google Scholar 
Kartzinel, T. R. et al. DNA metabarcoding illuminates dietary niche partitioning by African large herbivores. Proc. Natl. Acad. Sci. U.S.A. 112, 8019–8024 (2015).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Veldhuis, M. P. et al. Cross-boundary human impacts compromise the Serengeti-Mara ecosystem. Science 363, 1424–1428 (2019).ADS 
CAS 
PubMed 

Google Scholar 
Kavwele, C. M. et al. Non-local effects of human activity on the spatial distribution of migratory wildlife in Serengeti National Park, Tanzania. Ecol. Solut. Evid. 3, e12159 (2022).
Google Scholar 
Bijlsma, R. & Loeschcke, V. Environmental stress, adaptation and evolution: An overview. J. Evol. Biol. 18, 744–749 (2005).CAS 
PubMed 

Google Scholar 
Schmitt, M. H., Stears, K. & Shrader, A. M. Zebra reduce predation risk in mixed-species herds by eavesdropping on cues from giraffe. Behav. Ecol. 27, 1073–1077 (2016).
Google Scholar 
Preisser, E. L., Orrock, J. L. & Schmitz, O. J. Predator hunting mode and habitat domain alter nonconsmuptive effects in predator-prey interactions. Ecology 88, 2744–2751 (2007).PubMed 

Google Scholar 
Kiffner, C. et al. Long-term persistence of wildlife populations in a pastoral area. Ecol. Evol. 10, 10000–10016 (2020).PubMed 
PubMed Central 

Google Scholar 
Hopcraft, J. G. C. et al. Competition, predation, and migration: Individual choice patterns of Serengeti migrants captured by hierarchical models. Ecol. Monogr. 84, 355–372 (2014).
Google Scholar 
Fryxell, J. M. Forage quality and aggregation by large herbivores. Am. Nat. 138, 478–498 (1991).
Google Scholar 
Fitzgibbon, C. D. Mixed-species grouping in Thomson’s and Grant’s gazelles: The antipredator benefits. Anim. Behav. 39, 1116–1126 (1990).
Google Scholar 
Brown, J. S. & Kotler, B. P. Hazardous duty pay and the foraging cost of predation. Ecol. Lett. 7, 999–1014 (2004).
Google Scholar 
Stears, K. & Shrader, A. M. Increases in food availability can tempt oribi antelope into taking greater risks at both large and small spatial scales. Anim. Behav. 108, 155–164 (2015).
Google Scholar 
Creel, S. Toward a predictive theory of risk effects: Hypotheses for prey attributes and compensatory mortality. Ecology 92, 2190–2195 (2011).PubMed 

Google Scholar 
Périquet, S. et al. Effects of lions on behaviour and endocrine stress in plains zebras. Ethology 123, 667 (2017).
Google Scholar 
Stears, K., Schmitt, M. H., Wilmers, C. C. & Shrader, A. M. Mixed-species herding levels the landscape of fear. Proc. R. Soc. B Biol. Sci. 287, 20192555 (2020).
Google Scholar 
Schmitt, M. H., Stears, K., Wilmers, C. C. & Shrader, A. M. Determining the relative importance of dilution and detection for zebra foraging in mixed-species herds. Anim. Behav. 96, 151–158 (2014).
Google Scholar 
Meise, K., Franks, D. W. & Bro-Jørgensen, J. Alarm communication networks as a driver of community structure in African savannah herbivores. Ecol. Lett. 23, 293–304 (2020).PubMed 

Google Scholar 
Codron, D., Hofmann, R. R. & Clauss, M. Morphological and physiological adaptations for browsing and grazing. In The Ecology of Browsing and Grazing II (eds Gordon, I. J. & Prins, H. H. T.) 81–125 (Springer, 2019).
Google Scholar 
Odadi, W. O., Karachi, M. K., Abdulrazak, S. A. & Young, T. P. African wild ungulates compete with or facilitate cattle depending on season. Science 333, 1753–1755 (2011).ADS 
CAS 
PubMed 

Google Scholar 
Maestre, F. T., Callaway, R. M., Valladares, F. & Lortie, C. J. Refining the stress-gradient hypothesis for competition and facilitation in plant communities. J. Ecol. 97, 199–205 (2009).
Google Scholar 
de Jonge, M. M. J. et al. Conditional love? Co-occurrence patterns of drought-sensitive species in European grasslands are consistent with the stress-gradient hypothesis. Glob. Ecol. Biogeogr. 30, 1609–1620 (2021).PubMed 
PubMed Central 

Google Scholar 
Franks, D. W., Weiss, M. N., Silk, M. J., Perryman, R. J. Y. & Croft, D. P. Calculating effect sizes in animal social network analysis. Methods Ecol. Evol. 12, 33–41 (2021).
Google Scholar 
Estes, J. A. et al. Trophic downgrading of planet earth. Science 333, 301–306 (2011).ADS 
CAS 
PubMed 

Google Scholar 
Meise, K., Franks, D. W. & Bro-Jørgensen, J. Multiple adaptive and non-adaptive processes determine responsiveness to heterospecific alarm calls in African savannah herbivores. Proc. R. Soc. B Biol. Sci. 285, 20172676 (2018).
Google Scholar 
Blumstein, D. T., Bitton, A. & DaVeiga, J. How does the presence of predators influence the persistence of antipredator behavior? J. Theor. Biol. 239, 460–468 (2006).ADS 
MathSciNet 
PubMed 
MATH 

Google Scholar 
Riggio, J. et al. Lion populations may be declining in Africa but not as Bauer et al. suggest. Proc. Natl. Acad. Sci. 113, 201521506 (2015).
Google Scholar 
Bauer, H. et al. Lion (Panthera leo) populations are declining rapidly across Africa, except in intensively managed areas. Proc. Natl. Acad. Sci. 112, 14894–14899 (2015).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Pettorelli, N., Bro-Jørgensen, J., Durant, S. M., Blackburn, T. & Carbone, C. Energy availability and density estimates in African ungulates. Am. Nat. 173, 698–704 (2009).PubMed 

Google Scholar 
Haile, G. G. et al. Projected impacts of climate change on drought patterns over East Africa. Earth’s Future 8, 1–23 (2020).
Google Scholar 
Devine, A. P., McDonald, R. A., Quaife, T. & Maclean, I. M. D. Determinants of woody encroachment and cover in African savannas. Oecologia 183, 939–951 (2017).ADS 
PubMed 
PubMed Central 

Google Scholar 
Kiffner, C. et al. Long-term population dynamics in a multi-species assemblage of large herbivores in East Africa. Ecosphere 8, e02027 (2017).
Google Scholar 
Prins, H. H. T. & Loth, P. E. Rainfall patterns as background to plant phenology in northern Tanzania. J. Biogeogr. 15, 451–463 (1988).
Google Scholar 
Beattie, K., Olson, E. R., Kissui, B., Kirschbaum, A. & Kiffner, C. Predicting livestock depredation risk by African lions (Panthera leo) in a multi-use area of northern Tanzania. Eur. J. Wildl. Res. 66, 11 (2020).
Google Scholar 
Kasozi, H. & Montgomery, R. A. Variability in the estimation of ungulate group sizes complicates ecological inference. Ecol. Evol. 10, 6881–6889 (2020).PubMed 
PubMed Central 

Google Scholar 
USGS. MOD13Q1 v006 MODIS/Terra Vegetation Indices 16-Day L3 Global 250 m SIN Grid. 10.5067/MODIS/MOD13Q1.006 (2020).R Core Team. R: A Language and Environment for Statistical Computing. http://www.r-project.org/. Accessed January 02, 2022 (2021).Dice, L. R. Measures of the amount of ecologic association between species. Ecology 26, 297–302 (1945).
Google Scholar 
Croft, D. P., James, R. & Krause, J. Exploring Animal Social Networks (Princeton University Press, 2008).
Google Scholar 
Besag, J. & Clifford, P. Generalized Monte Carlo significance tests. Biometrika 76, 633–642 (1989).MathSciNet 
MATH 

Google Scholar 
Hayward, M. W. & Kerley, G. I. H. Prey preferences of the lion (Panthera leo). J. Zool. 267, 309–322 (2005).
Google Scholar 
Codron, D. et al. Diets of savanna ungulates from stable carbon isotope composition of faeces. J. Zool. 273, 21–29 (2007).
Google Scholar 
Kartzinel, T. R. & Pringle, R. M. Multiple dimensions of dietary diversity in large mammalian herbivores. J. Anim. Ecol. 89, 1482–1496 (2020).PubMed 

Google Scholar 
Prins, H. H. T. & Douglas-Hamilton, I. Stability in a multi-species assemblage of large herbivores in East Africa. Oecologia 83, 392–400 (1990).ADS 
CAS 
PubMed 

Google Scholar 
Tournier, E. et al. Differences in diet between six neighbouring groups of vervet monkeys. Ethology 120, 471–482 (2014).
Google Scholar 
Humphries, B. D., Ramesh, T. & Downs, C. T. Diet of black-backed jackals (Canis mesomelas) on farmlands in the KwaZulu-Natal Midlands, South Africa. Mammalia 80, 405–412 (2016).
Google Scholar  More

Thermogravimetric/differential thermogravimetry analyses of corn stoverThermogravimetric/Differential Thermogravimetry (TG/DTG) curves are shown in Fig. 2. The pyrolysis process of corn stover could be divided into three stages. The first stage was the dehydration stage, which occurred at approximately 55–125 °C, and the weight loss was mainly accounted for by water19. The second stage was the pyrolysis stage, which occurred at approximately 200–400 °C and mainly involved the decomposition of cellulose, hemicellulose and a small amount of lignin. This process involved the generation of CO and CO2 and the breaking of carbonaceous polymer bonds20. In addition, a shoulder peak in the range of 265 to 300 °C in the DTG diagram could be caused by side chain decomposition and glycosidic bond cleavage of xylan during the pyrolysis of corn stover21. The third stage was the carbonization stage, which occurred above 400 °C; this stage mainly involved the decomposition of lignin22,23. The carbonization process was relatively slow after 600 °C; this process was called the passive pyrolysis stage24. In general, the TG loss in the pyrolysis process of corn stover was mainly from the moisture in the biomass sample in the first stage. Hemicellulose and cellulose decomposition occurred in the second stage, and lignin decomposition occurred in the third stage25. In this experiment, the minimum pyrolysis temperature for the preparation of biochar was 400 °C. Therefore, the pyrolysis of biochar was relatively complete.Figure 2TG/DTG curves of corn stover.Full size imageCharacterization of biocharYield and specific surface area analysesThe yield and SBET are presented in Table 2. BC, BC-H and BC-OH represent the origin, acid-modified, and base-modified biochar, respectively. The yield of corn stover biochar exhibited a negative correlation with the temperature and decreased from 39.65 to 28.26% when the pyrolysis temperature increased from 400 to 700 °C. This phenomenon could have occurred due to the loss of more volatile substances and the thermal degradation of lignocellulose with increasing temperature, thus reducing the yield of biochar26,27. The SBET of the original biochar showed little difference below 700 °C but increased significantly at 700 °C. Combined with the SEM analysis (Fig. 3), at low temperatures, more ashes on the surface of biochar could block its pores so that the change in SBET was not obvious. At 700 °C, because the ash content significantly reduced and the pyrolysis was more sufficient, the pores of the biochar were more developed, and the SEBT significantly increased. The SBET of the acid/base-modified biochar increased with increasing temperature. The SBET of biochar was larger than that of the original biochar after acid and base modification at 400–600 °C. This phenomenon occurred because the porous structure of biochar was enhanced by acid and base modification28. Moreover, pickling removed most of the inorganic substances in biochar and reduced ash content, while alkali washing removed the tar on the surface of biochar to a certain extent29. However, at 700 °C, the SBET of biochar after acid/base modification was lower than that of the original biochar. Combined with the SEM (Fig. 4), the acid/base modification caused the nanopores of biochar to collapse into mesopores or macropores30. Therefore, the well-developed pore structure of the biochar prepared at 700 °C was destroyed by acid/base modification, resulting in a significant decrease in SBET.Table 2 Yield and SBET of different biochars.Full size tableFigure 3SEM (ZEISS) images of biochar at different pyrolysis temperatures: (a) C1, (b) C8, (c) C12, and (d) C16.Full size imageFigure 4SEM (OPTON) images of C16 biochar and its acid/base modification: (a) C16, (b) C16-H, and (c) C–OH.Full size imageScanning electron microscopy analysisThe C1, C8, C12 and C16 biochars had the highest Cd2+ removal rates at 400, 500, 600 and 700 °C, respectively. Therefore, these BCs were selected for SEM analysis. Figure 3 clearly showed that as the pyrolysis temperature increased from 400 to 700 °C, the pore structure of biochar became more developed, with a smaller pore size and more pores. Although there were numerous pores at 500 °C, the pores were not fully developed and were blocked inside. At 700 °C, the skeleton structure appeared, and the particle size of ash blocked in the pores decreased.By taking C16 biochar with the highest removal rate of Cd2+ as the research object, the changes in the biochar surface before and after modification were compared. C16-H and C16-OH represent acid-modified and base-modified biochar, respectively. After acid/base modification, the ash content on the surface of the biochar decreased, and the pore size increased (Fig. 4). Therefore, some skeleton structures could collapse after corrosion, which was consistent with the previous SBET results. Sun et al. discovered that citric acid-modified biochar would lead to micropore wall collapse and micropore loss, resulting in a reduction in SBET31. This finding was in agreement with the results of our study.Fourier transform infrared spectroscopy analysisThe FTIR spectra of biochar at different pyrolysis temperatures are presented in Fig. 5a.Figure 5FTIR spectra of corn stover biochar: (a) different pyrolysis temperatures and (b) different modification treatments.Full size imageAs the pyrolysis temperature increased from 300 to 700 °C, the absorption peak intensity showed a downwards trend. There was a remarkable decrease in features associated with stretch O–H (3400 cm−1)32. The vibration peaks of C–H (2924 cm−1) and C=O (1610 cm−1) decreased with increasing temperature, which could be due to the reduction in –CH2 and –CH3 groups of small molecules and the pyrolysis of C=O into gas or liquid byproducts at high temperatures33. In addition, the peak at 1435 cm−1 was identified as the vibration of C=C bonds belonging to the aromatic skeleton of biochar. A decrease in the absorbance peaks was found at 1115 cm−1, which corresponded to C–O–C bonds. The ratio of intensities for C=C/C=O (1550–1650 cm−1) and C–O–C (1115 cm−1) to the shoulder (1100–1200 cm−1) gradually decreased, and the loss of –OH at 3444 cm−1 indicated that the oxygen content in biochar reduced. The cellulose and wood components were dehydrated, and the degree of biochar condensation increased at higher temperatures. The bending vibration peaks of Ar–H at 856 and 877 cm−1 changed little at different temperatures, which showed that the aromatic rings were relatively stable below 700 °C34. Combined with the above analysis the condensation degree of biochar increased gradually above 400 °C35,36. In summary, as the pyrolysis temperature increased, the degree of aromatization of biochar improved, and the numbers of oxygen-containing functional groups decreased continuously.Figure 5b showed that after acid/base modification, the absorbance peaks at 3444 cm−1, 1610 cm−1 and 1115 cm−1 increased, indicating that the number of oxygen-containing functional groups increased. However, the stretching vibration peak of aromatic ring skeleton C=C (1435 cm−1) and the bending vibration peaks of Ar–H (856–877 cm−1) changed little. The number of functional groups of acid-modified biochar increased more than that of alkali-modified biochar. Mahdi et al. found that acid modification increased the number of functional groups in a study of biochar modification37. After acid/base modification, the number of oxygen-containing functional groups, such as hydroxyl and carboxyl groups, increased.Optimization of biocharFigure 6 illustrates that the removal rates of Cd2+ by corn stover biochar (original, acid-modified, and base-modified biochars) consistently increased with increasing pyrolysis temperature. The highest removal rate reached 95.79% at 700 °C. The removal rate decreased after modification, especially after pickling. The results showed that C16 biochar had the best removal effect on Cd2+.Figure 6Cd2+ removal rate of different biochars (BC: original biochar, BC-OH: alkali-modified biochar, and BC-H: acid-modified biochar).Full size imageIntuitive and variance analyses were employed to explore the influences of biochar preparation conditions on the removal rate of Cd2+.

1.

Intuitive analysis
The intuitive analysis of the orthogonal experiment is shown in Table 3 and Fig. 7. The pyrolysis temperature had the most significant influence on the removal of Cd2+, followed by the retention time and finally the heating rate. Therefore, the optimal conditions for biochar preparation were a pyrolysis temperature of 700 °C, a retention time of 2.5 h, and a heating rate of 5 °C/min.

2.

Variance analysis
Variance analysis showed that the effect of pyrolysis temperature on the removal rate of Cd2+ was very significant (Table 4). The effects of retention time and heating rate were not significant. This phenomenon was consistent with the conclusions obtained in the intuitive analysis.

Table 3 Intuitive analyses of influencing factors of biochar preparation.Full size tableFigure 7Intuitive analysis diagram of influencing factors for biochar preparation.Full size imageTable 4 Variance analysis.Full size tableAnalysis of adsorption mechanismThe SBET of the unmodified biochar did not change significantly with temperature, which indicated that SBET could potentially not be a critical factor for Cd2+ adsorption. Qi et al. obtained a similar conclusion when studying the adsorption of Cd2+ in water by chicken litter biochar38. In addition to SBET, the four primary mechanisms involved in the removal of heavy metal ions by biochar were as follows: (1) Ion exchange: the alkali or alkaline earth metals in biochar (K+, Ca2 +, Na+, and Mg2+) were the dominant cations in ion exchange39. (2) The complexation of oxygen-containing functional groups mainly included hydroxyl and carboxyl groups40. (3) Mineral precipitation: Cd2+ was precipitated by minerals on the surface of biochar to form Cd3(PO4)2 and CdCO341. Soluble cadmium precipitated with some anions released by biochar, such as CO32−, PO43− and OH−42,43. (4) π electron interaction: Cd2+ coordinated with the π electrons of C=C or C=O at low pyrolysis temperatures43,44. Biochar contains more aromatic structures at high pyrolysis temperatures, which could provide more π electrons. Therefore, the π electron interaction in adsorption of Cd2+ was effectively enhanced45.C1, C8, C12 and C16 were selected to study the adsorption mechanism. Related physicochemical properties are given in Table 5.Table 5 Physicochemical properties of biochar at different pyrolysis temperatures.Full size tableThe CEC of biochar gradually increased as the pyrolysis temperature increased, reaching a maximum at 600 °C and slightly decreasing at 700 °C. This phenomenon could have occurred because the crystalline minerals under high pyrolysis temperatures inhibited the exchange of cations on the surface of biochar with Cd2+ in aqueous solution46. Nevertheless, CEC did not change significantly with temperature; thus, CEC was not the main adsorption mechanism. With increasing pyrolysis temperature, the number of acidic functional groups decreased gradually, while the number of alkaline functional groups increased. The main functional groups used to remove Cd2+ were generally considered acidic oxygen-containing functional groups. However, the number of these functional groups decreased with increasing pyrolysis temperature, which weakened the complexation on the surface of the biochar. However, this result was contradictory to the results of Cd2+ adsorption. Therefore, the functional groups were not the main adsorption mechanism.To further explore the adsorption mechanism of Cd2+, the biochar before and after the adsorption of Cd2+ was characterized by XRD. As shown in Fig. 7a, C16-100Cd and C16-200Cd represented the biochar after Cd2+ adsorption when the concentrations of cadmium solution were 100 mg/l and 200 mg/l, respectively. The results showed that new peaks appeared at 30.275° and 36.546° after adsorption, corresponding to CdCO3. The spike at 29.454° was due to Cd(OH)2. Additionally, the intensity of the CdCO3 peak increased significantly from C16-100Cd to C16-200Cd, indicating that mineral precipitation occurred in adsorption. Liu et al. found similar results in a study on removing Cd2+ from water by blue algae biochar12. However, as the concentration of Cd2+ increased from 0 to 200 mg/L, the diffraction peak at 2θ = 29.454° first increased and then decreased. This because the peak position of CaCO3 at 2θ = 29.369° was very close to Cd(OH)2 at 2θ = 29.454°. At low concentrations, the production of Cd(OH)2 was greater than that of CdCO3. When the initial concentration of Cd2+ increased, more CO32− released by CaCO3 combined with Cd2+ to form CdCO3, resulting in a reduction in the diffraction peak.As presented in Fig. 8b, the peak intensities of CdCO3 and Cd(OH)2 gradually increase with increasing pyrolysis temperature. On the one hand, this phenomenon could be ascribed to the increase in the mineral content of biochar with increasing pyrolysis temperature. On the other hand, the pH value of biochar increased with increasing pyrolysis temperature. In this way, more OH− was released, thus forming more Cd(OH)2. Wang et al. obtained similar results42. Moreover, the peak intensity of KCl at 2θ = 28.347° decreased after adsorption, as shown in Fig. 8a, which indicated that ion exchange took part in adsorption.Figure 8XRD images: (a) before and after adsorption of Cd2+ on C16 biochar and (b) Cd2+ adsorption by biochar at different pyrolysis temperatures.Full size imageIn addition, the FTIR spectra showed that the number of functional groups, such as C=C and C=O, in biochar decreased with increasing pyrolysis temperature, leading to the weakening of cation–π interactions between Cd2+ and C=C and C=O. In contrast, due to the enhanced aromatization of functional groups on the surface of biochar, many lone pair electrons existed in the electron-rich domains of the graphene-like structure, which in turn enhanced the cation–π interactions. Harvey et al., based on the study of Cd2+ adsorption by plant biochar, concluded that the electron-rich domain bonding mechanism between Cd2+ and the graphene-like structure on the surface of biochar played a more significant role in biochar with a high degree of carbonization45. Therefore, π-electron interactions could play a dominant role in Cd2+ adsorption on high-temperature pyrolysis biochar. Moreover, the results showed that the number of alkaline functional groups increased while acidic functional groups decreased with the increase in pyrolysis temperature. It is generally believed that acidic functional groups could withdraw electrons, and basic functional groups could donate electrons47,48. The biochar with higher pyrolysis temperature contained more alkaline functional groups, which improved the electron donating ability of biochar and enhanced the cation–π electron effect.In summary, mineral precipitation and π electron coordination were the main mechanisms of removing Cd2+ from water by corn stover biochar. This phenomenon explained why the Cd2+ removal rate of acid/base–modified biochar decreased. After modification, the functional groups on the surface of biochar increased, but the inorganic minerals were removed. Pickling resulted in the loss of soluble minerals and alkaline functional groups on the surface of biochar, which was not conducive to adsorption49. After alkaline washing, more PO43−, CO32− and HCO3− were released, thereby reducing the mineral precipitation50,51. Since NaOH had a weaker destructive effect than HCl and introduced some OH−, alkaline washing had little effect on the removal rate of Cd2+.Adsorption isotherm and adsorption kineticsAdsorption isothermThe adsorption isotherms were fitted with Langmuir (Eq. 3) and Freundlich (Eq. 4) models, as shown in Fig. 9, and the fitting parameters are listed in Table 6.Figure 9Adsorption isotherm.Full size imageTable 6 Fitting parameters of the adsorption isotherm model.Full size tableThe Langmuir model (R2  > 0.963) was more suitable than the Freundlich model (R2  > 0.919), indicating that the adsorption sites of biochar were evenly distributed, and adsorption was mainly monolayer. Parameter KL reflected the difficulty of adsorption and was generally divided into four types: unfavourable (KL  > 1), favourable (0  More