Philippa Kaur

1.de Souza Dias, V., Pereira da Luz, M., Medero, G. M. & Tarley Ferreira Nascimento, D. An overview of hydropower reservoirs in Brazil: Current situation, future perspectives and impacts of climate change. Water 10, 592 (2018).
Google Scholar 
2.Patias, J., Zuquette, L. V. & Rodrigues-Carvalho, J. A. Piezometric variations in the basaltic massif beneath the Itaipu hydroelectric plant (Brazil/Paraguay border): Right Buttress Dam. Bull. Eng. Geol. Environ. 74, 207–231 (2015).CAS 

Google Scholar 
3.Agostinho, A. A. Pesquisas, monitoramento e manejo da fauna aquática em empreendimentos hidrelétricos. In Seminário Sobre Fauna Aquática E O Setor Elétrico Brasileiro 38–59 (Brasil, 1994).4.Makrakis, S., Gomes, L. C., Makrakis, M. C., Fernandez, D. R. & Pavanelli, C. S. The Canal da Piracema at Itaipu Dam as a fish pass system. Neotrop. Ichthyol. 5, 185–195 (2007).
Google Scholar 
5.Dos Reis, R. B., Frota, A., Depra, G. D. C., Ota, R. R. & Da Graca, W. J. Freshwater fishes from Paraná State, Brazil: An annotated list, with comments on biogeographic patterns, threats, and future perspectives. Zootaxa 4868, 451–494 (2020).
Google Scholar 
6.Becker, R. A., Sales, N. G., Santos, G. M., Santos, G. B. & Carvalho, D. C. DNA barcoding and morphological identification of neotropical ichthyoplankton from the Upper Paraná and São Francisco. J. Fish Biol. 87, 159–168 (2015).CAS 
PubMed 

Google Scholar 
7.Milan, D. T. et al. New 12S metabarcoding primers for enhanced Neotropical freshwater fish biodiversity assessment. Sci. Rep. 10, 1–12 (2020).ADS 

Google Scholar 
8.Agostinho, A. A., Pelicice, F. M. & Gomes, L. C. Dams and the fish fauna of the Neotropical region: Impacts and management related to diversity and fisheries. Braz. J. Biol. 68, 1119–1132 (2008).CAS 
PubMed 

Google Scholar 
9.Bonar, S. A., Hubert, W. A. & Willis, D. W. Standard methods for sampling North American freshwater fishes. American Fisheries Society, Bethesda, (USA, 2009).
Google Scholar 
10.Shaw, J. L. A. et al. Comparison of environmental DNA metabarcoding and conventional fish survey methods in a river system. Biol. Conserv. 197, 131–138 (2016).
Google Scholar 
11.Reis, R. E. et al. Fish biodiversity and conservation in South America. J. Fish Biol. 89, 12–47 (2016).CAS 
PubMed 

Google Scholar 
12.Baumgartner, G. et al. Peixes do baixo rio Iguaçu. (Eduem, 2012).13.Taberlet, P., Bonin, A., Coissac, E. & Zinger, L. Environmental DNA: For Biodiversity Research and Monitoring (Oxford University Press, 2018).
Google Scholar 
14.Taberlet, P., Coissac, E., Pompanon, F., Christian, B. & Willerslev, E. Towards next-generation biodiversity assessment using DNA metabarcoding. Mol Ecol 33, 2045–2050 (2012).
Google Scholar 
15.Ritter, C. D. et al. The pitfalls of biodiversity proxies: Differences in richness patterns of birds, trees and understudied diversity across Amazonia. Sci. Rep. 9, 1–3 (2019).CAS 

Google Scholar 
16.Sales, N. G. et al. Space-time dynamics in monitoring neotropical fish communities using eDNA metabarcoding. Sci. Total Environ. 754, 142096 (2021).ADS 
CAS 
PubMed 

Google Scholar 
17.Zinger, L. et al. Body size determines soil community assembly in a tropical forest. Mol. Ecol. 28, 528–543 (2019).CAS 
PubMed 

Google Scholar 
18.Baird, D. J. & Hajibabaei, M. Biomonitoring 2.0: A new paradigm in ecosystem assessment made possible by next-generation DNA sequencing. Mol. Ecol. 21, 2039–2044 (2012).PubMed 

Google Scholar 
19.Zinger, L. et al. Advances and prospects of environmental DNA in neotropical rainforests. Adv. Ecol. Res. 62, 331–373 (2020).
Google Scholar 
20.Cilleros, K. et al. Unlocking biodiversity and conservation studies in high-diversity environments using environmental DNA (eDNA): A test with Guianese freshwater fishes. Mol. Ecol. Resour. 19, 27–46 (2019).CAS 
PubMed 

Google Scholar 
21.Sales, N. G., Wangensteen, O. S., Carvalho, D. C. & Mariani, S. Influence of preservation methods, sample medium and sampling time on eDNA recovery in a neotropical river. Environ. DNA 119–130. https://doi.org/10.1002/edn3.14 (2020).Article 

Google Scholar 
22.Blaxter, M. et al. Defining operational taxonomic units using DNA barcode data. Philos. Trans. R. Soc. B Biol. Sci. 360(1462), 1935–1943. https://doi.org/10.1098/rstb.2005.1725 (2005).CAS 
Article 

Google Scholar 
23.Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581–583 (2016).CAS 
PubMed 
PubMed Central 

Google Scholar 
24.Edgar, R. C. UNOISE2: Improved error-correction for Illumina 16S and ITS amplicon sequencing. BioRxiv 81257 (2016).
25.Muha, T. P., Rodriguez-Barreto, D., O’Rorke, R., Garcia de Leaniz, C. & Consuegra, S. Using eDNA metabarcoding to monitor changes in fish community composition after barrier removal. Front. Ecol. Evol. 9, 28 (2021).
Google Scholar 
26.Kitano, T., Umetsu, K., Tian, W. & Osawa, M. Two universal primer sets for species identification among vertebrates. Int. J. Legal Med. 121, 423–427 (2007).PubMed 

Google Scholar 
27.Stoeckle, M. Y., Soboleva, L. & Charlop-Powers, Z. Aquatic environmental DNA detects seasonal fish abundance and habitat preference in an urban estuary. PLoS One 12, e0175186 (2017).PubMed 
PubMed Central 

Google Scholar 
28.Bylemans, J. et al. An environmental DNA-based method for monitoring spawning activity: A case study, using the endangered Macquarie perch (Macquaria australasica). Methods Ecol. Evol. 8, 646–655 (2017).
Google Scholar 
29.De Souza, L. S., Godwin, J. C., Renshaw, M. A. & Larson, E. Environmental DNA (eDNA) detection probability is influenced by seasonal activity of organisms. PLoS One 11, e0165273 (2016).PubMed 
PubMed Central 

Google Scholar 
30.Ritter, C. D. et al. Locality or habitat? Exploring predictors of biodiversity in Amazonia. Ecography (Cop.) 42, 321–333 (2019).
Google Scholar 
31.CFMV-Resolução no 1000 de 11 de maio de 2012—Dispõe sobre procedimentos e métodos de eutanásia em animais e dá outras providências. (2012).32.Britski, H. A., de Silimon, K. Z. S. & Lopes, B. S. Peixes do Pantanal: manual de identificação, ampl. Brasília, DF, Embrapa Informação Tecnológica (2007).33.Ota, R. R., Deprá, G. de C., Graça, W. J. da & Pavanelli, C. S. Peixes da planície de inundação do alto rio Paraná e áreas adjacentes: revised, annotated and updated. Neotrop. Ichthyol. 16(2). https://www.scielo.br/j/ni/a/tScwvm8JLhKnbxKjtBQLPBx/abstract/?lang=en (2018).34.Neris, N., Villalba, F., Kamada, D. & Viré, S. Guía de peces del Paraguay/Guide of fishes of Paraguay. Zamphiropolos, (Paraguay, 2010).
Google Scholar 
35.Pie, M. R. et al. Development of a real-time PCR assay for the detection of the golden mussel (Limnoperna fortunei, Mytilidae) in environmental samples. An. Acad. Bras. Cienc. 89, 1041–1045 (2017).CAS 
PubMed 

Google Scholar 
36.Miya, M. et al. MiFish, a set of universal PCR primers for metabarcoding environmental DNA from fishes: Detection of more than 230 subtropical marine species. R. Soc. Open Sci. 2, 150088 (2015).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
37.Boeger, W. A. et al. Testing a molecular protocol to monitor the presence of golden mussel larvae (Limnoperna fortunei) in plankton samples. J. Plankton Res. 29, 1015–1019 (2007).CAS 

Google Scholar 
38.Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 17, 10–12 (2011).
Google Scholar 
39.Van Rossum, G. & Drake, F. L. Python 3 References Manual. Scotts Valley CA: CreateSpace. (2009).40.R Core Team. R: the R project for statistical computing. 2019. https://www.r-project.org/ (accessed 30 Mar 2020).41.Edgar, R. C. UPARSE: Highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10, 996–998 (2013).CAS 
PubMed 

Google Scholar 
42.Edgar, R. C. & Flyvbjerg, H. Error filtering, pair assembly and error correction for next-generation sequencing reads. Bioinformatics 31, 3476–3482 (2015).CAS 
PubMed 

Google Scholar 
43.Camacho, C. et al. BLAST+: Architecture and applications. BMC Bioinform. 10, 421 (2009).
Google Scholar 
44.Team, Rs. RStudio: integrated development for R. RStudio, Inc., Boston, MA https://www.rstudio.com42, 84 (2015).45.Wickham, H. tidyverse: Easily Install and Load “Tidyverse” Packages (Version R package version 1.1. 1). (2017).46.Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer, 2016).MATH 

Google Scholar 
47.Tang, Y., Horikoshi, M. & Li, W. ggfortify: Unified interface to visualize statistical results of popular R packages. R J. 8, 474 (2016).
Google Scholar 
48.Auguie, B. & Antonov, A. gridExtra: Miscellaneous functions for “grid” graphics (Version 2.2. 1)[Computer software]. (2016).49.Kassambara, A. & Kassambara, M. A. Package ‘ggpubr’. (2020).50.Oksanen, J. et al. Vegan: Community ecology package. R package version 1.17-4. https://cran.r-project.org. Acesso em 23, 2010 (2010).51.McMurdie, P. J. & Holmes, S. Waste not, want not: Why rarefying microbiome data is inadmissible. PLoS Comput. Biol. 10, e1003531 (2014).ADS 
PubMed 
PubMed Central 

Google Scholar 
52.Jost, L. Entropy and diversity. Oikos 113, 363–375 (2006).
Google Scholar 
53.Marcon, E., Herault, B. & Marcon, M. E. Package ‘entropart’. (2021).54.Mächler, E., Walser, J.-C. & Altermatt, F. Decision making and best practices for taxonomy-free eDNA metabarcoding in biomonitoring using Hill numbers. BioRxiv (2020).55.McMurdie, P. J. & Holmes, S. phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE 8, e61217 (2013).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
56.Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).PubMed 
PubMed Central 

Google Scholar 
57.León, A., Reyes, J., Burriel, V. & Valverde, F. Data quality problems when integrating genomic information. In International Conference on Conceptual Modeling 173–182 (Springer, 2016).58.Callahan, B. J., McMurdie, P. J. & Holmes, S. P. Exact sequence variants should replace operational taxonomic units in marker-gene data analysis. ISME J. 11, 2639–2643 (2017).PubMed 
PubMed Central 

Google Scholar 
59.Stahlhut, J. K. et al. DNA barcoding reveals diversity of hymenoptera and the dominance of parasitoids in a sub-arctic environment. BMC Ecol. 13, 2 (2013).PubMed 
PubMed Central 

Google Scholar 
60.Gillet, B. et al. Direct fishing and eDNA metabarcoding for biomonitoring during a 3-year survey significantly improves number of fish detected around a South East Asian reservoir. PLoS One 13, e0208592 (2018).PubMed 
PubMed Central 

Google Scholar 
61.Barrett, M. et al. Living planet report 2018: Aiming higher. WWF. Available at: https://www.globallandscapesforum.org/publication/living-planet-report-2018-aiming-higher/ (2018).62.Díaz, S. M. et al. The global assessment report on biodiversity and ecosystem services: Summary for policy makers. Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. 56, (2019).63.Dudgeon, D. Asian river fishes in the Anthropocene: Threats and conservation challenges in an era of rapid environmental change. J. Fish Biol. 79, 1487–1524 (2011).CAS 
PubMed 

Google Scholar 
64.Dudgeon, D. Multiple threats imperil freshwater biodiversity in the Anthropocene. Curr. Biol. 29, R960–R967 (2019).CAS 
PubMed 

Google Scholar 
65.He, F. et al. Disappearing giants: A review of threats to freshwater megafauna. Wiley Interdiscip. Rev. Water 4, e1208 (2017).
Google Scholar 
66.Agostinho, A. A., Thomaz, S. M. & Gomes, L. C. Threats for biodiversity in the floodplain of the Upper Paraná River: Effects of hydrological regulation by dams. (2018). Int. J. Ecohydrol. Hydrobiol Warsaw. 4(3), 267–280 (2004).
Google Scholar 
67.Santana, M. L., Carvalho, F. R. & Teresa, F. B. Broad and fine-scale threats on threatened Brazilian freshwater fish: Variability across hydrographic regions and taxonomic groups. Biota Neotrop. 21 (2). https://www.scielo.br/j/bn/a/YqFbWSy5vbfHy3QK9kNpdKp/?format=html&lang=en (2021).68.Matthews, W. J. Patterns in Freshwater Fish Ecology. (Springer Science & Business Media, 2012).69.de Oliveira Bueno, E., Alves, G. J. & Mello, C. R. Hydroelectricity water footprint in Parana hydrograph region, Brazil. Renew. Energy 162, 596–612 (2020).
Google Scholar 
70.Camacho Guerreiro, A. I., Amadio, S. A., Fabre, N. N. & da Silva Batista, V. Exploring the effect of strong hydrological droughts and floods on populational parameters of Semaprochilodus insignis (Actinopterygii: Prochilodontidae) from the Central Amazonia. Environ. Dev. Sustain. 23, 3338–3348 (2021).
Google Scholar 
71.Jespersen, H., Rasmussen, G. & Pedersen, S. Severity of summer drought as predictor for smolt recruitment in migratory brown trout (Salmo trutta). Ecol. Freshw. Fish 30, 115–124 (2021).
Google Scholar 
72.Pool, T. K., Grenouillet, G. & Villéger, S. Species contribute differently to the taxonomic, functional, and phylogenetic alpha and beta diversity of freshwater fish communities. Divers. Distrib. 20, 1235–1244 (2014).
Google Scholar 
73.de Oliveira, E. F., Goulart, E. & Minte-Vera, C. V. Fish diversity along spatial gradients in the Itaipu Reservoir, Paraná, Brazil. Braz. J. Biol. 64, 447–458 (2004).CAS 
PubMed 

Google Scholar 
74.Daga, V. S. et al. Homogenization dynamics of the fish assemblages in Neotropical reservoirs: Comparing the roles of introduced species and their vectors. Hydrobiologia 746, 327–347 (2015).
Google Scholar 
75.Vitule, J. R. S. Introdução de peixes em ecossistemas continentais brasileiros: revisão, comentários e sugestões de ações contra o inimigo quase invisível. Neotrop. Biol. Conserv. 4, 111–122 (2009).
Google Scholar 
76.Mariac, C. et al. Species‐level ichthyoplankton dynamics for 97 fishes in two major river basins of the Amazon using quantitative metabarcoding. Mol. Ecol. https://onlinelibrary.wiley.com/action/showCitFormats?doi=10.1111%2Fmec.15944 (2021).77.Jackman, J. M. et al. eDNA in a bottleneck: Obstacles to fish metabarcoding studies in megadiverse freshwater systems. Environ. DNA 3, 837–849 (2021).
Google Scholar 
78.Bessey, C. et al. Maximizing fish detection with eDNA metabarcoding. Environ. DNA 2, 493–504 (2020).
Google Scholar 
79.Evans, N. T. et al. Fish community assessment with eDNA metabarcoding: Effects of sampling design and bioinformatic filtering. Can. J. Fish. Aquat. Sci. 74, 1362–1374 (2017).CAS 

Google Scholar 
80.Prodan, A. et al. Comparing bioinformatic pipelines for microbial 16S rRNA amplicon sequencing. PLoS One 15, e0227434 (2020).CAS 
PubMed 
PubMed Central 

Google Scholar 
81.Edgar, R. C. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461 (2010).CAS 

Google Scholar 
82.Elbrecht, V. & Leese, F. Can DNA-based ecosystem assessments quantify species abundance? Testing primer bias and biomass—sequence relationships with an innovative metabarcoding protocol. PLoS ONE 10, e0130324 (2015).PubMed 
PubMed Central 

Google Scholar 
83.Pawluczyk, M. et al. Quantitative evaluation of bias in PCR amplification and next-generation sequencing derived from metabarcoding samples. Anal. Bioanal. Chem. 407, 1841–1848 (2015).CAS 
PubMed 

Google Scholar 
84.Holman, L. E., Chng, Y. & Rius, M. How does eDNA decay affect metabarcoding experiments? Environ. DNA https://onlinelibrary.wiley.com/action/showCitFormats?doi=10.1002%2Fedn3.201 (2021).85.Edgar, R. C. UNCROSS2: identification of cross-talk in 16S rRNA OTU tables. BioRxiv 400762 (2018).86.MacArthur, R. H. Geographical Ecology: Patterns in the Distribution of Species. (Princeton University Press, 1984).87.Leray, M. & Knowlton, N. Random sampling causes the low reproducibility of rare eukaryotic OTUs in Illumina COI metabarcoding. PeerJ 5, e3006 (2017).PubMed 
PubMed Central 

Google Scholar 
88.Team, Q. D. QGIS geographic information system. Open Source Geospatial Found. Proj. Versão 2, (2015). More

Microbial community samplingHippo gut microbiomeWe characterized the microbial communities in the hippo gut by collecting ten samples of fresh hippo feces adjacent to four hippo pools early in the morning (prior to desiccation by the sun) in September 2017. We collected feces from different pools and locations adjacent to the pool to include the feces of different individuals so we could estimate the similarity of the gut microbiome among individuals and across the landscape. The four hippo pools are sufficiently far apart that there was likely no intermixing of hippos among them.Each individual hippo feces sample was gently homogenized by hand and then the liquid was gently squeezed from the coarse particulate organic matter. A portion of the liquid (approximately 10 mL) was vacuum filtered through a Supor polysulfone membrane (0.2-µm pore size; Pall, Port Washington, NY, USA). After approximately 10 mL had filtered through and the filter was dry, 15 mL of RNALater was gently poured onto the filter and allowed to contact the collected biomass on the filter for 15 min before being removed by filtration. The filter was stored dry in a sterile petri dish and transferred to a refrigerator within several hours, then to a − 20 °C freezer for storage within several days.During the July 2016 survey of hippo pools, we collected an additional two samples of fresh hippo feces near a high-subsidy hippo pool and filtered approximately 10 mL of the liquid portion after homogenization as detailed above. The filter was then folded twice to preserve the biomass on the filter and stored in 14 mL of RNALater.Aquatic ecosystemWe characterized the microbial communities in the water column of hippo pools across a gradient of hippo subsidy (July 2016, N = 12 pools). We collected samples from the upstream, downstream, surface, and bottom of both pools containing hippos and pools that lacked hippos. Subsamples were also analyzed for biogeochemical variables (details provided below). We also collected water samples in four of the high-subsidy hippo pools every 2–3 days starting immediately after a flushing event until the next flushing event (August and September 2017, Supplementary Fig. S1)23. The number of hippos, discharge and volume for each pool are presented in Dutton et al (2020)23.We sampled the aquatic microbial community and biogeochemical variables along a longitudinal transect down both the Mara and Talek rivers (Supplementary Fig. S1, Supplementary Table S2). For the Mara River, we sampled an approximately 100-km transect along a gradient of hippo numbers (N = 10 locations, from 0 to ~ 4000 hippos). For the Talek River, we sampled an approximately 30-km transect to the confluence with the Mara (N = 8 locations, from 0 to 700 hippos). Mara River sites 9 and 10 are downstream of the confluence with the Talek River. Water samples were collected from each site in a well-mixed flowing section away from any hippo pools.Aquatic microbial samples were collected by filtering water samples through a Supor polysulfone filter (0.2-µm pore size; Pall, Port Washington, NY, USA) and then preserving the filter in RNALater Stabilization Solution (Ambion, Inc., Austin, TX, USA). In 2017, the filters were preserved with RNA Later and then frozen for analysis.Mesocosm experimentWe collected river water from the Mara River upstream of the distribution of hippos and placed it in 45 1-L bottles in a large water basin covered by a dark tarp to help regulate temperature and prevent algal production. Bottles were randomly assigned to the control, bacteria, and bacteria + virus treatments. We collected fresh hippo feces from multiple locations adjacent to the Mara River. After homogenization, half of the hippo feces was sterilized in a pressure cooker, which testing confirmed had similar sterilization results as an autoclave53 (see Supplementary Materials). Five grams of sterilized hippo feces was placed into each bottle to provide an organic matter substrate without viable bacteria or viruses. The unsterilized hippo feces was expressed, and the resulting liquid was filtered through 0.7-µm GF/F filters (0.7-µm pore size; Whatman, GE Healthcare Life Sciences, Pittsburgh, PA, USA) and 0.2-µm Supor filters to physically separate the bacteria (on the filter papers) from the viruses (in the filtrate). Half the filtrate was then sterilized with a UV light treatment (Supplementary Fig. S4). The UV light treatment did not significantly alter DOC quality (see Supplementary Materials).We prepared 15 bottles for each of three treatments—control, bacteria, and bacteria + virus—as follows: Control Unfiltered river water, 5 g wet weight sterilized hippo feces, and two blank Supor filters; Bacteria Unfiltered river water, 5 g wet weight sterilized hippo feces, two Supor filters containing bacteria, and 4 mL sterilized filtrate; Virus Unfiltered river water, 5 g wet weight sterilized hippo feces, two Supor filters containing bacteria, and 4 mL unsterilized filtrate containing viruses.We conducted the experiment for 27 days from September to October 2017. We terminated the experiment after 27 days because we were trying to replicate the microbial communities in hippo pools as best as we could and the hippo pools rarely go more than 1 month before they are flushed out by a flood25. Initial microbial samples of the river water, hippo feces bacteria and hippo fecal liquid filtrate were taken on day 0, and three replicate samples per treatment were destructively sampled on day 3, 9, 15, 21, and 27. During each time step, the microbial communities were sampled using the methods detailed above, and chemical analyses were done on the water samples as described below. We also measured chlorophyll a, dissolved oxygen, temperature, conductivity, total dissolved solids, turbidity, and pH with a Manta2 water quality sonde (Eureka Water Probes, Austin, TX, USA).Microbial community characterizationWe used 16S rRNA sequencing to characterize the active microbial communities. We extracted both DNA and RNA from our preserved samples, then used RNA to synthesize cDNA to represent the “active” microbial community and the total DNA in the sample to represent the “total” microbes present, including those that may not be actively replicating54. Due to the continual loading of hippo feces into pools and the long half-life of DNA, we would expect there to be significant quantities of microbial DNA derived from hippo feces within the pools. However, there would be less accumulation of RNA because of RNA’s shorter half-life. The active communities identified through this RNA-based approach are the ones that would potentially contribute to ecosystem function55 as indicated by the protein synthesis potential, although relationships between activity and rRNA concentrations in individual taxa within mixed communities can vary56. Nevertheless, this method provides an overall characterization of the microbial community’s potential activity.We used the Qiagen RNeasy Powerwater Kit (Qiagen, Hilden, Germany) to extract the DNA and RNA from the material on the filter using a slightly modified manufacturer’s protocol to allow for the extraction of both DNA and RNA. After extraction, we split the total extracted volume (100 µL per sample) into two groups. We treated one group with the DNase Max Kit (Qiagen, Hilden, Germany) to remove all DNA and serve as the RNA group of samples.We used the RNA group of samples to synthesize cDNA using the SuperScript III First Strand Synthesis Kit (Invitrogen, Carlsbad, CA, USA). DNA and cDNA were quantified using the PicoGreen dsDNA Assay Kit (Molecular Probes, Eugene, OR, USA) then normalized to 5 ng/µL. Amplicon library preparation was done using a dual-index paired-end approach57. We amplified the V4 region of the 16S rRNA gene using dual-index primers (F515/R805) and AccuPrime Pfx SuperMix (Invitrogen, Carlsbad, CA, USA) in duplicate for each sample using the manufacturer’s recommended thermocycling routine.Samples were then pooled, purified and normalized using the SequelPrep Normalization Plate Kit (Invitrogen, Carlsbad, CA, USA). Barcoded amplicon libraries were then sequenced at the Yale Center for Genome Analysis (New Haven, CT, USA) using an Illumina Miseq v2 reagent kit (Illumina, San Diego, CA, USA) to generate 2 × 250 base pair paired-end reads.Sampling took place in 2016 and 2017 and involved two separate sequencing runs. The first sequencing run included negative controls and a mock community (D6306, Zymo Research, Irvin, CA, USA). The second sequencing run included negative controls, a mock community (D6306), and a single E. coli strain. In both runs, the mock community and single E. coli strain were well reconstructed from the sequences, and there was minimal contamination in the negative controls, mock community and E. coli strain.From those two sequencing campaigns, we received over 2 million raw sequences from the first campaign and over 7 million raw sequences for the second campaign. For the microbial community analyses, only samples collected and sequenced during the same campaign are analyzed together to prevent preservation or sequencing biases. However, samples within the two separate campaigns were preserved and sequenced using identical methods with only a minor modification (mentioned above) to increase the preservation of genetic material.We de-multiplexed sequenced reads then removed barcodes, indexes, and primers using QIIME258. We used DADA2 with a standard workflow in R59 to infer exact sequence variants (ESV) for each sample60. We assigned taxonomy using a naïve Bayesian classifier and the SILVA training set v. 128 database61,62. We removed potential contamination in samples from both campaigns by using the statistical technique in the R package, decontam63. We used Phyloseq to characterize, ordinate, and compare microbial communities64 with their standard workflow59.Chemical analysesAll water samples collected in the field and in the experiment were analyzed for dissolved ferrous iron (Fe(II)), hydrogen sulfide (H2S), dissolved organic carbon (DOC), inorganic nutrients, major ions, dissolved gases, and biochemical oxygen demand following the standard methods provided in detail in Dutton et al (2020)23.Statistical analysesWe computed all statistical analyses in the R 4.1.1 statistical language in RStudio 2021.09.0 using α = 0.05 to determine significance65,66. Error bars in the figures represent standard deviation of the means. All data and R code for the statistics and data treatments are provided in the Mendeley Data Online Repository67.We used the Bray–Curtis dissimilarity matrix followed by ordination with NMDS to examine differences between individual hippo gut microbiomes; between low-, medium-, and high-subsidy hippo pools; and between a gradient of hippo pools and the environment. We used a CCA to test for the influence of biogeochemical drivers on microbial community composition using biogeochemical data that were previously published but collected concurrently with this study23. We constrained the CCA ordination by soluble reactive phosphorus, nitrate, methane, BOD, and sulfate, which were all previously shown to be important drivers in the variation between pools23. We used PERMANOVA and PERMDISP to test for significant differences between groups68.
We compared aquatic microbial communities from the bottom of high-subsidy hippo pools (N = 15), from hippo feces (N = 10, the hippo gut microbiome) and upstream of high-subsidy hippo pools (N = 15, free of hippo gut microbiome influence) using the Bray–Curtis dissimilarity matrix on the relative abundances for the active aquatic microbial communities collected from the different sample types followed by ordination with NMDS. 95% confidence ellipses were generated. We then determined the active taxa that were shared between the hippo gut microbiome (hippo feces) and the bottom of the high-subsidy hippo pools and not present in the upstream samples from high-subsidy hippo pools.We used LEfSe to calculate the differential abundance of microbial taxa between upstream (N = 14), downstream (N = 16), at the surface (N = 17) and at the bottom (N = 14) of hippo pools and calculated their effect size69. We then calculated the correlation of microbial taxa to the measured biogeochemistry using Pearson’s correlation coefficient with a false discovery rate corrected p-value in the microeco R package70.
We used SourceTracker to quantify the contribution of the hippo gut, upstream waters, or unknown sources to the active aquatic microbial communities in the bottom waters of three of the high-subsidy hippo pools between flushing flows71. We also used the Bray–Curtis dissimilarity matrix followed by ordination with NMDS to examine changes in the active aquatic microbial communities in one of the high subsidy hippo pools through time after flushing flows.For the experiment, we calculated the Bray–Curtis dissimilatory matrix followed by ordination with NMDS for the active aquatic microbial communities over time in each of the three experimental treatments. We used SourceTracker to determine the proportion of the active aquatic microbial community in each treatment that originated from the hippo gut, the river water, or unknown sources71. We analyzed the biogeochemical differences among experimental treatments by fitting a linear mixed effects model for each of the biogeochemical variables throughout the experiment with the nlme package in R72. We fit the model with the restricted maximum likelihood method and a continuous autoregressive temporal correlation structure with sample day as the repeated factor. Treatment and time were fixed effects and individual bottles were treated as random effects. We conducted a pairwise post-hoc test with an ANOVA and the emmeans package in R to estimate marginal means with a Tukey adjusted p-value for multiple comparisons73,74. More

MPs abundanceIn Table 1 MP abundance (mean value ± standard deviation) values are presented by shape, size range, color and polymer type categories for each sampling site. MP were found in all analyzed salt samples including pellets, fibers, fragments, films and lines (Fig. 3). MP total abundance values per site ranged from 74.7 to 136.7 particles kg−1 in the following order of increasing abundance: S3  black (17%)  > blue (15%)  > green and transparent (10% each)  > pink (6%)  > colorless (5%). In terms of size, most particles were in the category 500–1000 µm, except for S3 (1000–5000 µm) (Table 1). The distribution of MP particles based on size range was: 500–1000 µm (40%)  > 1000–5000 µm (34%)  > 250–500 µm (26%). For salts from the Atlantic and the Pacific Ocean, originating from Brazil, the United Kingdom, and the USA, Kim et al.12 reported a higher abundance of MP in size range 100–1000 µm while sizes in the range 100–5000 µm were reported for salt samples from the Black Sea. Seth and Shriwastay20 found that 80% of fibers found in salt samples from the Indian Sea were smaller than 2000 μm in length. MP size range differences among the various studies are suggested to depend on the degree of weathering for a given environment30, different climatic conditions such as wind, rain, temperature, salinity, and waves influencing size range composition. Also, for runoff, rivers, and atmospheric fallout transportation, smaller MP size ranges can be expected to be associated with a longer range from the initial plastic sources31,32,33. Nevertheless, more detailed information about MP polymer/color features within the size ranges are needed to achieve stronger conclusions about potential long/short-range sources.Figure 6Microplastics abundance (particles kg−1) by color in sea salt samples from stations S1 to S8.Full size imageFigure 7Microplastics abundance (particles kg−1) by size range in sea salt samples from stations S1 to S8.Full size imageMP polymer compositionFour types of polymer, namely polypropylene (PP), polystyrene (PS), polyethylene (PE), and polyethylene terephthalate (PET), were identified with FT-MIR-NIR (Supplementary Figure S1). These results are in accordance with those reported for salt samples in other studies worldwide (Table 1). These polymer types are widely used in daily life products, packaging, single-use plastics, and clothes, contributing to plastic pollution worldwide21. PET presented the highest contribution at all sampling sites, at ~ 48%, whereas PS was found to be least, at ~ 15% (Fig. 8, Table 1). Iñiguez et al.34 also reported PET predominance (83.3%) in Spanish table salt samples. PET predominance could be explained by its high density (1.30 g cm−3), making particles prone to sedimentation during the salt crystallization process19. PE (0.94 g cm−3), PP (0.90 g cm−3), and PS (1.05 cm−3) presented lower or similar densities to seawater (~ 1.02 g cm−3), making these more prone to flotation and possible loss due to wind during desiccation.Figure 8Microplastics abundances (particles kg−1) by polymer composition in sea salt samples from stations S1 to S8.Full size imageRisks assessmentDuring degradation, MP tends to emit monomers and different types of additives, these having the potential to cause harm to ecological systems and health18, 35. Results for the polymeric risks indices are presented in Fig. 9. According to polymer risk classification, all salts samples showed low risks, being similar to the entire study area. To date, none of the published studies have applied chemometric models in evaluating MP pollution in salts, posing difficulties when comparing our results. Information on the hazards of MP from ingestion to human health is still highly unclear. Other than exposure, the destiny and transit of ingested MP in the human body, including intestinal digestion and biliary discharge, have not been determined in previous research and remained largely unclear36. Some studies conducted impact assessments based on in vitro models37,38. However, whether the exposure concentrations used in such studies indicate the MP consumed and collected in humans is inconclusive. Previous studies found that toxicity, oxidative stress, and inflammation could result from MP exposure, including immune disruption and neurotoxicity effects, among others39. Therefore, an immediate effort is required to assess the health consequences of these MP when they reach the human body.Figure 9Polymeric risk indices for MP types in salts from stations S1 to S8.Full size image More

AffiliationsDepartment of Ocean Sciences, University of California, Santa Cruz, CA, USARachel A. FosterDepartment of Biogeochemistry, Max Planck Institute for Marine Microbiology, Bremen, GermanyRachel A. Foster, Daniela Tienken, Sten Littmann & Marcel M. M. KuypersDepartment of Ecology, Environment and Plant Sciences, Stockholm University, Stockholm, SwedenRachel A. FosterDepartment of Geosciences, Swedish Museum of Natural History, Stockholm, SwedenMartin J. WhitehouseDepartment of Oceanography, University of Hawai’i at Mānoa, Honolulu, HI, USAAngelicque E. WhiteAuthorsRachel A. FosterDaniela TienkenSten LittmannMartin J. WhitehouseMarcel M. M. KuypersAngelicque E. WhiteCorresponding authorCorrespondence to
Rachel A. Foster. More

Data for all analysed radionuclides are presented in the “Supplementary Material”. Cryoconite samples were collected on Nalli Glacier (Supplementary Fig. S1) on Sept. 25, 2017 (samples 1701–1714) and on Sept. 10, 2018 (samples 1801–1814) at 28 spots (Fig. 2, Supplementary Table S1). Gamma spectrometric analysis of samples showed the presence of anthropogenic radionuclides 137Cs, 241Am, and 207Bi. All quoted radioactivity values were recalculated for the sampling date, except those for 241Am since the concentration of the parent 241Pu isotope is unknown. However, for this isotope, the correction for decay is negligible. The activity of 137Cs reached 8093 (± 69) Bq kg−1 of dry weight, that of 241Am reached 58.3 (± 2.3) Bq kg−1 and that of 207Bi reached 6.3 (± 0.6) Bq kg−1. The natural radionuclides 210Pb and 7Be were also present in all samples. The activity of 210Pb varied in the range of 1280–9750 Bq kg−1. In addition, in the investigated samples, a significant amount of short-lived cosmogenic radionuclide 7Be was found, whose specific activity reached 2418 (± 76) Bq kg−1 (Fig. 3, Supplementary Table S2). To evaluate the contribution of atmospheric components to the total 210Pb activity, 226Ra activity was determined and found to be 17–27 Bq kg−1 (Supplementary Table S2). Based on the 210Pb/226Ra ratio, we conclude that more than 98% of 210Pb was of atmospheric provenance.Figure 2Location of sampling points on Nalli Glacier. A—137Cs activity zone  95%) of corresponding rocks and numerous outcrops likely promoted entrapment of these elements into explosion clouds, and their subsequent precipitation with radionuclides. This feature of the geological structure of the area explains the extremely high enrichment of surface waters in elements such as Zn, Pb, Sr, Ni, As, Cr, Co, Se, Te, Cd, W, Cu, Sb, and Sn; for many of them, the excess reaches 10-fold with respect to the Clrake values51. This hypothesis is supported by obvious correlations between the concentrations of Bi, Ag, Sn, Sb, Pb, Cd, W, and Cu and the activity of anthropogenic radionuclides 137Cs, 241Am and 207Bi. This relationship is obviously related to the simultaneous release of elements and radionuclides from the contaminated ice layer and their entrapment in cryoconite holes. More

Download PDF

For the past five years, I’ve studied oysters — a commercially and environmentally important species in southeast England. My research is very practical: I help to solve problems by working with oyster growers (known locally as oystermen), regulators and other community members. Resulting papers are evidence of work I’ve already done.Most oysters in this area are a non-native species (Crassostrea giga). Locally, it’s well established and has been since the 1960s, but allowing it to spread to nearby estuary systems has been controversial: there are concerns that it could become an invasive species.Working with aquaculture producers, I help to guide efforts to restore the native oyster (Ostrea edulis), populations of which declined owing to overfishing, habitat destruction, pollution and disease. Crassostrea giga oysters have provided enough income for oyster growers to spend time and effort restoring the local species. We’ve done some cool things, including creating one of the largest coastal marine conservation zones in the United Kingdom — more than 284 square kilometres — and all for an unseen brown invertebrate that lacks the charisma of a dolphin.This picture is from a typical day in the field. During high tides, we go out in a boat to take sonar readings to map potential oyster habitats; at low tide, we put on waders and go out on the mud flats to look for juvenile oysters. We focus our conservation efforts on spots where juvenile oysters are already trying to get established.Amazingly, these filter feeders don’t require feeding by humans, and they clean the water as they grow. Bivalve aquaculture such as this has become a cornerstone of the ‘blue economy’ — using marine resources sustainably for economic growth while preserving ocean health. It will take more work to determine how the balance can be reached, but oysters will be part of that conversation.

Nature 600, 182 (2021)
doi: https://doi.org/10.1038/d41586-021-03573-5

Related Articles

How AI is helping the natural sciences

Breeding the sweetest biofuels in the business

Using sound to explore events of the Universe

Subjects

Conservation biology

Agriculture

Latest on:

Agriculture

Even organic pesticides spur change in the wildlife next door
Research Highlight 17 NOV 21

Tracking the origin of Transeurasian languages
News & Views 10 NOV 21

New rules will make UK gene-edited crop research easier
News 30 SEP 21

Jobs

Full Professor in Microbial Biotechnology

University of Groningen (RUG)
Gorningen, Netherlands

Research Faculty Position Cardiovascular Research Center Massachusetts General Hospital

Massachusetts General Hospital
Charlestown, MA, United States

Higher Scientific Officer – Biophysical assay development and protein production

Institute of Cancer Research (ICR)
London, United Kingdom

Group Leaders

John Innes Centre (JIC)
Norwich, United Kingdom More

1.Agetsuma, N. Foraging strategies of Yakushima Macaques (Macaca-fuscata Yakui). Int. J. Primatol. 16, 595–609. https://doi.org/10.1007/bf02735283 (1995).Article 

Google Scholar 
2.Hill, D. A. Seasonal variation in the feeding behavior and diet of Japanese macaques (Macaca fuscata yakui) in lowland forest of Yakushima. Am. J. Primatol. 43, 305–322 (1997).CAS 
Article 

Google Scholar 
3.Otani, Y. et al. Factors influencing riverine utilization patterns in two sympatric macaques. Sci. Rep. https://doi.org/10.1038/s41598-020-79259-1 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
4.Maruhashi, T. Feeding behavior and diet of the Japanese monkey Macaca-fuscata-yakui on Yakushima Island, Japan. Primates 21, 141–160. https://doi.org/10.1007/bf02374030 (1980).Article 

Google Scholar 
5.Rothman, J. M., Raubenheimer, D., Bryer, M. A. H., Takahashi, M. & Gilbert, C. C. Nutritional contributions of insects to primate diets: Implications for primate evolution. J. Hum. Evol. 71, 59–69. https://doi.org/10.1016/j.jhevol.2014.02.016 (2014).Article 
PubMed 

Google Scholar 
6.Hanya, G. et al. Not only annual food abundance but also fallback food quality determines the Japanese macaque density: evidence from seasonal variations in home range size. Primates 47, 275–278. https://doi.org/10.1007/s10329-005-1076-2 (2006).ADS 
Article 
PubMed 

Google Scholar 
7.Nakagawa, N. Determinants of the dramatic seasonal changes in the intake of energy and protein by Japanese monkeys in a cool temperate forest. Am. J. Primatol. 41, 267–288. https://doi.org/10.1002/(sici)1098-2345(1997)41:4%3c267::aid-ajp1%3e3.0.co;2-v (1997).CAS 
Article 
PubMed 

Google Scholar 
8.Tsuji, Y., Ito, T. Y., Wada, K. & Watanabe, K. Spatial patterns in the diet of the Japanese macaque Macaca fuscata and their environmental determinants. Mammal Rev. 45, 227–238. https://doi.org/10.1111/mam.12045 (2015).Article 

Google Scholar 
9.Wada, K. & Tokida, E. Habitat utlization by wintering Japanese monkeys Macaca fuscata-fuscata in Shiga Heights Japan. Primates 22, 330–348. https://doi.org/10.1007/bf02381574 (1981).Article 

Google Scholar 
10.Suzuki, A. An ecological study of wild Japanese monkeys in snowy area focused on their food habits. Primates 6, 31–71 (1965).Article 

Google Scholar 
11.Izawa, K. & Nishida, T. Monkeys living in the northern limits of their distribution. Primates 4, 67–88 (1963).Article 

Google Scholar 
12.Enari, H. & Sakamaki-Enari, H. Influence of heavy snow on the feeding behavior of Japanese Macaques (Macaca Fuscata) in Northern Japan. Am. J. Primatol. 75, 534–544. https://doi.org/10.1002/ajp.22128 (2013).Article 
PubMed 

Google Scholar 
13.Agetsuma, N. Dietary selection by Yakushima macaques (Macaca-fustcata Yakui): the influence of food availability and temperature. Int. J. Primatol. 16, 611–627. https://doi.org/10.1007/bf02735284 (1995).Article 

Google Scholar 
14.Agetsuma, N. & Nakagawa, N. Effects of habitat differences on feeding behaviors of Japanese monkeys: comparison between Yakushima and Kinkazan. Primates 39, 275–289. https://doi.org/10.1007/bf02573077 (1998).Article 

Google Scholar 
15.Izumiyama, S. In: High Altitude Primates, Developments in Primatology: Progress and Prospects Vol. 44 (ed N.B. Grow et al.) 153–181 (Springer, New York, 2014).16.Go, M. Seasonal changes in food resource distribution and feeding sites selected by Japanese macaques on Koshima Islet, Japan. Primates 51, 149–158. https://doi.org/10.1007/s10329-009-0179-5 (2010).Article 
PubMed 

Google Scholar 
17.Hanya, G. Diet of a Japanese macaque troop in the coniferous forest of Yakushima. Int. J. Primatol. 25, 55–71. https://doi.org/10.1023/b:ijop.0000014645.78610.32 (2004).Article 

Google Scholar 
18.Sakamaki, H., Enari, H., Aoi, T. & Kunisaki, T. Winter food abundance for Japanese monkeys in differently aged Japanese cedar plantations in snowy regions. Mammal Study 36, 1–10. https://doi.org/10.3106/041.036.0101 (2011).Article 

Google Scholar 
19.Enari, H. In: High Altitude Primates. Developments in Primatology, Progress and Prospects (eds N Grow, S Gursky-Doyen, & Krzton A) (Springer, New York, 2014).20.Tsuji, Y. & Nakagawa, N. Monkeys of Japan: A Mammalogical Studies of Japanese Macaques (University of Tokyo Press, 2017).
Google Scholar 
21.Suzuki, S., Hill, D. A., Maruhashi, T. & Tsukuhara, T. Frog and Lizard-eating behaviour of wild Japanese Macaques in Yakushima, Japan. Primates 31, 421–426 (1990).Article 

Google Scholar 
22.Watanabe, K. Fish: a new addition to the diet of Japanese macaques on Koshima Island. Folia Primatol. 52, 124–131. https://doi.org/10.1159/000156391 (1989).CAS 
Article 

Google Scholar 
23.Leca, J. B., Gunst, N., Watanabe, K. & Huffman, M. A. A new case of fish-eating in Japanese macaques: Implications for social constraints on the diffusion of feeding innovation. Am. J. Primatol. 69, 821–828. https://doi.org/10.1002/ajp.20401 (2007).Article 
PubMed 

Google Scholar 
24.Stewart, A. M. E., Gordon, C. H., Wich, S. A., Schroor, P. & Meijaard, E. Fishing in Macaca fascicularis: a rarely observed innovative behavior. Int. J. Primatol. 29, 543–548. https://doi.org/10.1007/s10764-007-9176-y (2008).Article 

Google Scholar 
25.Hamilton, W. J. & Tilson, R. L. Fishing Baboons at desert waterholes. Am. J. Primatol. 8, 255–257. https://doi.org/10.1002/ajp.1350080308 (1985).Article 
PubMed 

Google Scholar 
26.Tamura, M. Extractive foraging on hard-shelled walnuts and variation of feeding techniques in wild Japanese macaques (Macada fuscata). Am. J. Primatol. 82, e23130 (2020).Article 

Google Scholar 
27.Iwamoto, T. A bioeconomic study on a provisioned troop at a Japanese monkeys Macada fuscata-fuscata at Koshima Islet Miyazaki. Primates 15, 241–262. https://doi.org/10.1007/bf01742286 (1974).Article 

Google Scholar 
28.Tsuji, Y. & Takatsuki, S. Effects of a typhoon on foraging behavior and foraging success of Macaca fuscata on Kinkazan Island, Northern Japan. Int. J. Primatol. 29, 1203–1217. https://doi.org/10.1007/s10764-008-9293-2 (2008).Article 

Google Scholar 
29.Gumert, M. D. & Malaivijitnond, S. Marine prey processed with stone tools by burmese long-tailed macaques (Macaca fascicularis aurea) in intertidal habitats. Am. J. Phys. Anthropol. 149, 447–457. https://doi.org/10.1002/ajpa.22143 (2012).Article 
PubMed 

Google Scholar 
30.Tan, A., Tan, S. H., Vyas, D., Malaivijitnond, S. & Gumert, M. D. There is more than one way to crack an oyster: identifying variation in burmese long-tailed Macaque (Macaca fascicularis aurea) stone-tool use. PLoS ONE https://doi.org/10.1371/journal.pone.0124733 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
31.Urabe, M. The present distribution and issues regarding the control of the exotic snail Potamopyrgus antipodarum in Japan. Jpn. J. Limnol. 68, 491–496 (2007).Article 

Google Scholar 
32.Hamada, K. T. Y. & Urabe, M. Survey of mitochondrial DNA haplotypes of Potamopyrgus antipodarum (Caenogastropoda: Hydrobiidae) introduced into Japan. Limnology 14, 223–228 (2013).Article 

Google Scholar 
33.Izumiyama, S., Mochizuki, T. & Shiraishi, T. Troop size, home range area and seasonal range use of the Japanese macaque in the Northern Japan Alps. Ecol. Res. 18, 465–474. https://doi.org/10.1046/j.1440-1703.2003.00570.x (2003).Article 

Google Scholar 
34.Milner, A. M., Docherty, C., Windsor, F. M. & Tojo, K. Macroinvertebrate communities in streams with contrasting water sources in the Japanese Alps. Ecol. Evol. 10, 7812–7825. https://doi.org/10.1002/ece3.6507 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
35.Vestheim, H. & Jarman, S. N. Blocking primers to enhance PCR amplification of rare sequences in mixed samples: a case study on prey DNA in Antarctic krill stomachs. Front. Zool. https://doi.org/10.1186/1742-9994-5-12 (2008).Article 
PubMed 
PubMed Central 

Google Scholar 
36.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. https://doi.org/10.1186/1742-9994-10-34 (2013).Article 
PubMed 
PubMed Central 

Google Scholar 
37.Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. 2011 17, 3. doi:https://doi.org/10.14806/ej.17.1.200 (2011).38.Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581. https://doi.org/10.1038/nmeth.3869 (2016).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
39.Ratnasingham, S. & Hebert, P. D. N. BOLD: the barcode of life data system (www.barcodinglife.org). Mol. Ecol. Notes 7, 355–364. doi:https://doi.org/10.1111/j.1471-8286.2007.01678.x (2007).40.Wang, Q., Garrity, G. M., Tiedje, J. M. & Cole, J. R. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73, 5261–5267. https://doi.org/10.1128/aem.00062-07 (2007).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
41.McMurdie, P. J. & Holmes, S. Phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE https://doi.org/10.1371/journal.pone.0061217 (2013).Article 
PubMed 
PubMed Central 

Google Scholar 
42.Wickham, H. ggplot2: elegant graphics for data analysis 2nd edn. (Springer, 2016).Book 

Google Scholar  More

As a starting point, we compare the vulnerability of four districts in Lyon, Paris (France), Firenze (Italy) and New York (US). These cities were chosen as emblematic of different topologies, resulting from different historical urban layering. The historic center of Firenze (panel b in Fig. 1) is mainly characterized by a dense urban fabric with a medieval signature of narrow and winding streets24. In Paris (panel c), Haussman’s renovation plan at the end of the 19th century supplemented the North–South and East–West ancient crossroad by a second network of concentric large avenues25. The rectilinear grid of Manhattan, New York, originates from 181126,27 and extends along the spine of Manhattan island (panel d). Despite the significant difference in size, a similar regular pattern is found in the modern urban area of Lyon (panel a), developed in the second half of the 19th century. In the insets of Fig. 1, we report for each city a polar histogram of the orientation of the streets. Although greater variability is observable for the orientation of the streets in the urban areas of Firenze (panel b) and Paris (panel c), two main orthogonal axes are found in the spatial structure of each city.The urban networks analysed in this work were delimited in order to be large enough to include the distinctive patterns of these four cities. The edges of the areas were traced along physical boundaries (e.g., rivers, parks, railways, large avenues) which act as elements of discontinuity in the dispersion process. Where not possible, the break was forced along wide streets.We promptly computed vulnerability maps for the selected urban areas by means of the centrality metric we derived in20 and recall in the “Methods” section. The nodes with the highest centrality values (V) are the most vulnerable as they correspond to the best spreading locations in the urban fabric. The spreading potential of a node is evaluated based on the extent of the area that is contaminated when the release takes place in this same node.We report in Fig. 1 the vulnerability maps of the four urban areas for the indicative scenario of a wind blowing at an angle (phi =45^circ). In the insets of Fig. 1, the wind direction is indicated with a red arrow. Given the different orientation and structure of the street networks, (phi) is defined as a clockwise angle with respect to the main axis of the city, which is identified as the longest bar in the polar histogram of street orientation.To extend the analysis to multiple meteorological scenarios, we estimated the vulnerability of each node (seen as a spreading source) for eight different wind directions ((phi =0^circ), (45^circ), (90^circ), (135^circ), (180^circ), (225^circ), (270^circ), (315^circ)). In this way, for each city, we obtained an extended dataset of vulnerability values that we represent in a compact way by means of a cumulative distribution function, as shown in Fig. 2a. The intercept of the cdf represents the nodes with null vulnerability. These are mostly located along the physical edges of the domain where the pollutant gas is blown away by the wind without affecting other streets. Where the delimitation of the network is forced (for example on the sides of central park as regards Manhattan), the interruption of the propagation, in the vulnerability model, is also constrained. This does not result in any artificial effect when the boundary is located upwind with respect to the network (propagation carries on from the boundary towards the considered urban area). On the other hand, when the boundary is downwind, vulnerability can be there underestimated. Considering the multiple wind directions simulated and the small number of nodes belonging to these edges (1% of the total number of network nodes), this effect has been calculated negligible to the purposes of this work.According to the mean values (vertical dashed lines) of the distributions reported in Fig. 2, New York is the most vulnerable city on average, while Firenze is the most protected. The vulnerability of New York and Lyon are the most sensitive to changes in wind direction, as shown by Fig. 2.b, where a polar histogram reports the mean vulnerability for each city for the eight directions of the approaching wind. In general, the spreading potential is more effective when the wind is oblique ((phi =45^circ ,) (135^circ), (225^circ), and (315^circ)) to the main orthogonal axes of the street network, as evidenced by the higher vulnerability observed for the dark gray sectors of Fig. 2b. We also notice that vulnerability for parallel ((phi =0^circ ,) (phi =180^circ)) and perpendicular ((phi =90^circ ,) (phi =270^circ)) wind directions is quite similar. This seems counterintuitive as previous studies (e.g.,28,29) have reported that a perpendicular wind is much more unfavorable for the dispersion of pollutants in a street. In this regard, we underline that (phi) is here defined with respect to the main axis of the city, so for (phi =90^circ) not all streets will be perpendicular to the wind direction. For example, in the regular network of Manhattan we expect the number of perpendicular streets to be similar to that of parallel streets, when (phi =90^circ).Figure 1Vulnerability maps for (a) Lyon, (b) Firenze, (c) Paris, and (d) New York for a wind direction of (45^circ) with respect to the main axis of the urban fabric. The polar histograms in the insets report the distribution of street orientation, while the red arrows represent the wind direction with respect to the street network. Panels a1–d1 show the urban pattern in a rectangular area of 0.5 km(^2) (reported in panels a–d) for the cities of Lyon, Firenze, Paris, and New York, respectively. Background images made with QGIS 2.18 (https://qgis.org).Full size imageFigure 2Vulnerability distribution for different cities and wind directions. (a) Cdf of node vulnerability for the different cities under eight different wind directions. The mean vulnerability is shown as a dashed line and reported numerically together with the standard deviation (in parentheses). (b) Mean vulnerability of city networks for each wind direction. Colors blue, yellow, green and magenta correspond to the urban networks of Lyon, Firenze, Paris and New York, respectively.Full size imageThe reasons for the different resilience of cities (and their patterning) to gas propagation are embedded in the centrality metric adopted to compute urban vulnerability. The key factors for node vulnerability can then be analytically recognized in the metric definition (Eqs. 4–5 in “Methods”): the highest vulnerabilities are achieved when the set of reachable nodes ((mathcal {V})) from the source node is large, and the paths connecting the source and the reachable nodes ((d_{sr})) are short, i.e. the propagation cost ((omega)) along the paths is minimal. In other words, the spots in a city (i.e. nodes in a network) with the highest spreading potential are those from which a toxic plume can reach many other locations with significant concentration. Going beyond the vulnerability results, we aim here to decompose the aforementioned elaborate and meaningful quantities (the set of reachable nodes, the shortest paths, the propagation cost) in elementary properties of the urban area in order to link the vulnerability of a city to its tangible characteristics.We start by disassembling the propagation cost associated to each street. Given a source node, a pollutant plume will propagate along the streets downwind the node. The propagation cost of each street (Eq. 4) describes the decay of concentration that the plume undergoes when it propagates along the street. Neglecting physico-chemical transformations, this cost depends on the transport processes within the streets and is a function of two dimensionless quantities: a geometrical ratio between the length (l) and height (h) of the street canyon, and a dynamic ratio between the exchange rate of pollutants towards the atmosphere above roof level (v) and the advective velocity along the longitudinal (u) axis of the street. According to30 and31, these two velocities can be parametrized as a function of the external wind intensity, the cosine ((theta)) of the angle between the wind direction and the orientation of the street, the geometry of the street canyon (its length l, height h and width w) and the aerodynamic roughness of building walls. As detailed in the Methods, the dependence of the propagation cost on the external wind intensity disappears as both velocities u and v scale linearly with it. Assuming constant aerodynamic resistance of the surfaces, the parameters l, h, w, (theta), remain the relevant building blocks for the propagation cost along a street.We underline that the parametrizations adopted here for the transport mechanisms in a street are based on the up-to-date literature and are currently employed in operational models (see the Methods section for mode details). Any refinements to this transport model may be included in the future. In this case, the cost associated to each street may depend on additional parameters that, however, we expect to be of second-order importance to those listed above.While pollutant transport in a single street canyon (i.e. the propagation cost) has been easily broken down into its basic elements, the information enclosed in the shortest paths ((d_{sr})) and in the set of reachable nodes from the source ((mathcal {V})) is much more challenging to trace back to evident properties of the city. These quantities depend on the sequence of streets that must be traveled to connect a source node to the surrounding nodes, i.e. on the way the streets are interconnected. The information is thus primarily topological. However, we point out that the interconnectivity of the network is not frozen, but dynamic, as it is given by the reaction of the urban structure to the direction of the external wind. In fact, the links of the street network are directed according to the orientation of the approaching wind. Moreover, the connectivity between the nodes is limited by the decay of the concentration along the streets. Although a target node may be reached from the source node by means of a path across the network, the two nodes may not actually be connected by a propagation path as the pollutant concentration may vanish along the path. For these reasons, traditional descriptors of network topology cannot be applied directly to describe the topological component of the vulnerability. Instead, we have to look for tailored and simple indicators that can express the wind-driven interconnectivity of the street network and the reachability potential between the nodes.Focusing on a node as spreading source, we infer that the number of links in its downwind area gives a first estimate of the potential for a release in the node to affect many other locations in the network. To delimit this downwind area, we adopt the concept of n-hop neighborhood32,33. Two nodes are n hops apart if it is possible to reach the target node from the source node by traveling n links. We identify the downwind area of the source node as the subnetwork composed by the nodes that are reachable from the source via at most n hops along the directed links. We propose the number of links in this neighborhood (k) as a suitable measure of reachability from the node. This reachability depends upon three features: (i) the local structure of the street network, (ii) the direction of the wind, and (iii) the topological distance n. This latter parameter is intuitively correlated to the intensity of the release. More precisely, it depends on the ratio between the magnitude of the toxic release at the source and the threshold value for pollutant concentration to be significant. In this work, n is taken as constant and its value is obtained from an optimization analysis detailed in the “Methods” section (Fig. 6) .Once the (n-hop) neighborhood of a node is delimited, the number of links k is not exhaustive in giving information about the properties of the paths connecting the source to the other nodes of the neighborhood. For the same k, different structures of the neighborhood can take place (see Fig. 6b), with consequent different outcomes for the propagation process that we are breaking down to basic components. The higher the number of links outgoing each node of the neighborhood, the higher the potential concentration for the k links, as they are topologically closer to the source. This feature can be accounted for by means of a simple branching index (b) for the node neighborhood, defined in Eq. 8 as the average outdegree for the nodes belonging to the neighborhood34.The disassembling analysis presented above suggests that the spreading potential of a node, and thus its vulnerability, mainly depends on the topological parameters k and b and on the geometrical characteristics of its neighborhood, i.e. L, H, W, (Theta), where the capital letters are used to indicate the local average (over the n-hop neighborhood) for the length (l), height (h), width (w), and orientation ((theta)) of the street canyons.In adopting averaged geometrical properties, we are assuming that these characteristics are rather homogeneous in the surroundings of a node. While the height, width and length of the street canyons are actually quite uniform on a local scale, especially in European city centers, the same does not apply to the orientation of the streets. The streets of a neighborhood intersect each other at different angles (e.g., at (90^circ) in grid plans), and the intensity of the wind in the streets changes strongly with their orientation. Low wind streets act as bottlenecks in the propagation paths, thus strongly influencing the spreading dynamics. For this reason, the standard deviation of street orientation in the neighborhood ((sigma)) is expected to be an additional topological index of node vulnerability.To assess whether the identified parameters are valuable basic elements of node vulnerability, we perform a regression analysis adopting a simple (but versatile) non linear model of the form:$$begin{aligned} V_{pred}=alpha L^beta H^gamma W^delta Theta ^epsilon k^zeta b^eta (1-sigma )^lambda . end{aligned}$$
(1)
We estimate the coefficients (alpha) to (lambda) by means of a nonlinear least square technique (namely the fitnlm function in Matlab) that minimizes the sum of the squares of the residuals between the predicted vulnerability (V_{pred}) and the vulnerability V obtained from the centrality metric (Eq. 5 in “Methods”). The regression is performed considering all the scenarios presented in this study: four different urban networks and eight different wind directions. The p-values for the coefficients (alpha) to (lambda) tend to zero, indicating that the relationships between the independent variables and the observations (V) are statistically significant. Note that in Eq. (1) we adopt (1-sigma) as predictor, instead of (sigma), to avoid null entries, as (sigma) takes value in [0 1). To explain the reason for this range for (sigma), we point out that the angle between the wind direction and the street axis is defined in [(-90^circ) (90^circ)]. As a consequence, the cosine ((theta)) of the angle varies in [0 1] and the standard deviation of (theta) (i.e. (sigma)) varies in [0 1).The scatter plot in Fig. 3 compares (V_{pred}) against V. Points correspond to the nodes of the four urban networks in the eight wind scenarios. The figure suggests that 80% of the spreading capacity (V) of a spot in a city can be grasped from the basic geometrical and topological characteristics of its neighborhood. To identify the most influential parameters in the regression, we evaluate the gain in the coefficient of determination (R^2) as they are progressively included in the model (red circles in the inset). The quantities are entered in order to optimize (R^2) at each addition. Alternatively, the role of each parameter can be evaluated adopting the concept of unique contribution (triangles in the inset), i.e. the loss in the coefficient of determination induced by the exclusion of the parameter from the model35. Both analyses reveal k and (sigma) as the main indicators for the vulnerability of a node. Actually, more than 60% of the total variance (inset in Fig. 3) is explained by these two parameters, unveiling the crucial role of topology in governing the dynamics of pollutants in urban areas. The effect of the geometrical properties (L, H, W, (Theta)) of the street canyons is secondary. Among these, the contribution of the building height (H) is the most remarkable as its contribution, combined with that of the two topological parameters k and (sigma), brings the correlation to almost its maximum value.Given these results, it is enlightening to show some tangible examples of how the three simple indicators k, (sigma) and H dominate urban vulnerability. We wonder which of these properties determine the distinct vulnerability of neighboring areas belonging to the same district, and which ones differentiate the resilience of cities with a different urban history.Figure 3Correlation of node vulnerability with basic geometrical and topological parameters of the street network. Color (blue to red) is associated to point density. Left y-axis of inset: trend of the coefficient of determination (R^2) as the urban indicators are progressively included in the model. Right y-axis of inset: unique contribution of the indicators.Full size imageFigure 4a shows the spatial distribution of the key parameters k, (sigma) and H and of node vulnerability, for Manhattan and a wind direction (phi =45^circ). In panel b, high street reachability (k) is observed in the central part of Midtown, in the heart of Downtown, and near Wall Street. An homogeneous distribution in the orientation of the streets with respect to the incident wind (low values for (sigma) and thus high (1-sigma)) is especially found in Midtown (panel c). Finally, in panel d, high-buildings (H) distinguish the Financial District and East Midtown. A perfect match between the four layers is not expected as vulnerability is given by the synergistic contribution of the different parameters. However, in line with the results of the regression model shown in Fig. 3, a positive correlation is observable between the most vulnerable areas (circled areas comprising the nodes with highest V in panel a of Fig. 4) and those with the highest values for the three indicators. In these areas, high buildings inhibit the vertical exchange of pollutants between the streets and the atmosphere above, as largely discussed in literature (see e.g.36,37). This inhibition limits the concentration decay along the propagation paths and facilitates large-scale contamination. Moreover, the great number (k) of streets topologically close to the node increase the impact of the release. The effect of (sigma) is significant especially for the vulnerability of Midtown. Here, since (phi =45^circ) and the street network is regular, (theta) (the cosine of the wind-street angle) is almost the same for all the streets. Therefore, the standard deviation of (theta) ((sigma)) is low and the predictor (1-sigma) is high. Physically, this means that the external wind approaches all the streets with almost the same angle. As a consequence, the intensity of the longitudinal wind in the streets is similar (the street aspect ratio is also similar) and the propagation takes place equally along both the dominant and lateral segments of the street network38, thus favoring the spread over large areas. Although high values of H and (1-sigma) can be detected in the North-East corner of Midtown too, here the vulnerability is mitigated by a higher discontinuity in the urban pattern (low k). This feature, together with the great overlapping of red areas in panel a with those in panel b, evidences the key role of street reachability (k) in the heterogeneity of vulnerability between areas of the same urban district.Figure 4Street network of Midtown and Downtown Manhattan. Node color is associated to node vulnerability (V), and to its key indicators: street reachability (k), inhomogeneity in street orientation ((sigma)), and average height of buildings in the node neighborhood (H).Full size imageFrom these observations, we move to a broader view and investigate the structural fragility of a city as a whole. In Fig. 5a–c, we report the probability density function (pdf) of the key parameters k, (1-sigma) and H. For each city, the pdf is calculated over all the network nodes and for the different wind directions. So, each pdf is representative of eight different networks for the same urban area. In panel a, the distributions for the four cities are quite similar but the tails of the pdfs highlight that the highest values for street reachability (k) occur in the street networks of Lyon and New York. The homogeneity in street orientation with respect to the wind (panel b), expressed by (1-sigma), exhibits a bimodal distribution and a slightly higher mean for the regular street network of New York. The two peaks are associated to distinctive wind scenarios, as will be discussed below. Also in this case, the observation of the tails of the pdfs reveals that high values of the vulnerability indicator are more probable in Lyon and New York. Finally, the distribution of building height (panel c) presents the most marked difference between the considered architectures, with high-rise buildings contributing to the heavy pdf tail of Manhattan. Comparing these results with those in Fig. 2a, Manhattan’s greatest vulnerability appears to be due to the greater depth of the urban canyons (high H) and the greater homogeneity, on average, in wind-street orientation (high (1-sigma)). Conversely, the medieval structure of Firenze, with higher heterogeneity in street orientation (low (1-sigma)) and low buildings (low H), enhances street ventilation and hinders propagation over long distances. Moreover, the tails of the pdfs for k and (1-sigma) reveal the role of topology in the higher variability of vulnerability values (given by the standard deviation of the pdfs in Fig. 2) for the street networks of Lyon and Manhattan.After discussing the behavior of the single parameters, we assess the synergistic contribution of the three quantities. To this aim, we define a simple correlation index (rho =widehat{k} cdot (1-widehat{sigma }) cdot widehat{H}), where the hat denotes a min-max normalization of the parameters, i.e. the range of values of each parameter is rescaled in [0, 1]. For the urban areas of Manhattan, Lyon, Paris, and Firenze, (rho) gives 0.039, 0.017, 0.012, and 0.011, respectively. This ranking complies with the ranking inferable in Fig. 2 for the average vulnerability of the cities. This result confirms that vulnerability occurs when the three parameters are correlated, as already evidenced in Fig. 4.To make the picture even more fascinating, it is worth noting that the role of topology, shown above as key, is dynamic as it varies according to the direction of the wind impacting the urban fabric. In panels d to f of Fig. 5, the pdfs of k, (1-sigma) and H are distinguished for four wind directions ((phi =0^circ), (45^circ), (90^circ) and (135^circ)). For each angle, the statistics are calculated over the examined cities, together. Although wind orientation alters the direction of the network links, and thus the delimitation of the n-hop neighboring area of each node, street reachability (panel d) and building height (panel f) remain statistically invariant for the different wind directions, suggesting a rather isotropic structure of the urban fabric. On the other hand, the variability in street orientation with respect to the wind (panel e) presents two distinctive trends for wind directions aligned with or oblique to the main axes of the street network. To explain this behavior, we refer to the simple case of a grid-like urban plan, like Manhattan’s plan. When (phi =0^circ) or (90^circ), (theta) (the cosine of the angle between the street and the wind direction) mostly switches between 0 (for the streets aligned with the wind) and 1 (for the orthogonal streets), resulting in a high standard deviation over the neighborhood (low (1-sigma)). When (phi =45^circ) or (135^circ), instead, the incident angle (theta) mainly takes intermediate values, leading to higher values for (1-sigma). This distinctive behavior is clearly detectable in the two peaks that we have observed in panel b for the regular grid of Lyon and New York. The left peak of the bimodal distribution corresponds to the scenarios with aligned wind directions, while the right peak occurs for oblique wind directions over the city. A more irregular street pattern in Firenze and Paris adds random contributions to the way the wind approaches the street, thus altering this bimodal shape. Going back to panel e, the greater homogeneity in wind-street orientation (higher (1-sigma)) for (phi =45^circ) or (135^circ) gives insights into the higher vulnerability found for the scenarios with these wind directions in almost all cities (dark gray sectors in Fig. 2b). This result is confirmed by the correlation ((rho)) between the three rescaled parameters ((widehat{k}), (1-widehat{sigma }), (widehat{H})). The correlation (rho) is estimated separately for the different wind directions, but considering the nodes from the four urban areas together. For oblique wind directions, (rho) is about twice ((rho =0.035)) the value found for the aligned wind directions ((rho =0.018)).Figure 5Probability density function of the key parameters k, (1-sigma), H. In the first row, each curve refers to a city and includes vulnerability data from eight different wind directions. In the second row, each curve corresponds to a specific wind direction and includes vulnerability data from the four cities, together.Full size image More