Ecology

1.Marden, J. H. & Waddington, K. D. Floral choices by honeybees in relation to the relative distances to flowers. Physiol. Entomol. 6, 431–435 (1981).Article 

Google Scholar 
2.Waddington, K. D., Allen, T. & Heinrich, B. Floral preferences of bumblebees (Bombus edwardsii) in relation to intermittent versus continuous rewards. Anim. Behav. 29, 779–784 (1981).Article 

Google Scholar 
3.Bauer, A. A., Clayton, M. K. & Brunet, J. Floral traits influencing plant attractiveness to three bee species: consequences for plant reproductive success. Am. J. Bot. 104, 772–781 (2017).PubMed 
Article 

Google Scholar 
4.Bradshaw, H. D. & Schemske, D. W. Allele substitution at a flower colour locus produces a pollinator shift in monkeyflowers. Nature 426, 176–178 (2003).ADS 
CAS 
PubMed 
Article 

Google Scholar 
5.Klahre, U. et al. Pollinator choice in petunia depends on two major genetic loci for floral scent production. Curr. Biol. 21, 730–739 (2011).CAS 
PubMed 
Article 

Google Scholar 
6.Muth, F., Papaj, D. R. & Leonard, A. S. Bees remember flowers for more than one reason: pollen mediates associative learning. Anim. Behav. 111, 93–100 (2016).Article 

Google Scholar 
7.Brunet, J., Thairu, M. W., Henss, J. M., Link, R. I. & Kluever, J. A. The effects of flower, floral display, and reward sizes on bumblebee foraging behavior when pollen is the reward and plants are dichogamous. Int. J. Plant Sci. 176, 811–819 (2015).Article 

Google Scholar 
8.Nicholls, E. & De Ibarra, N. H. Bees associate colour cues with differences in pollen rewards. J. Exp. Biol. 217, 2783–2788 (2014).PubMed 
Article 

Google Scholar 
9.Thairu, M. W. & Brunet, J. The role of pollinators in maintaining variation in flower colour in the Rocky Mountain columbine, Aquilegia coerulea. Ann. Bot. 115, 971–979 (2015).PubMed 
PubMed Central 
Article 

Google Scholar 
10.Ishii, H. S. Floral display size influences subsequent plant choice by bumble bees. Funct. Ecol. 20, 233–238 (2006).Article 

Google Scholar 
11.Mitchell, R. J., Karron, J. D., Holmquist, K. G. & Bell, J. M. The influence of Mimulus ringens floral display size on pollinator visitation patterns. Funct. Ecol. 18, 116–124 (2004).Article 

Google Scholar 
12.Makino, T. T. & Sakai, S. Experience changes pollinator responses to floral display size: from size-based to reward-based foraging. Funct. Ecol. 21, 854–863 (2007).Article 

Google Scholar 
13.Osborne, J. L. et al. A landscape-scale study of bumble bee foraging range and constancy, using harmonic radar. J. Appl. Ecol. 36, 519–533 (1999).Article 

Google Scholar 
14.Osborne, J. L. & Williams, I. H. Site constancy of bumble bees in an experimentally patchy habitat. Agric. Ecosyst. Environ. 83, 129–141 (2001).Article 

Google Scholar 
15.Saville, N. M., Dramstad, W. E., Fry, G. L. A. & Corbet, S. A. Bumblebee movement in a fragmented agricultural landscape. Agric. Ecosyst. Environ. 61, 145–154 (1997).Article 

Google Scholar 
16.Ogilvie, J. E. & Thomson, J. D. Site fidelity by bees drives pollination facilitation in sequentially blooming plant species. Ecology 97, 1442–1451 (2016).PubMed 
Article 
PubMed Central 

Google Scholar 
17.Cresswell, J. E. & Osborne, J. L. The effect of patch size and separation on bumblebbe foraging in oilseed rape: implications for gene flow. J. Appl. Ecol. 41, 539–546 (2004).Article 

Google Scholar 
18.Ohashi, K. & Thomson, J. D. Trapline foraging by pollinators: its ontogeny, economics and possible consequences for plants. Ann. Bot. 103, 1365–1378 (2009).PubMed 
PubMed Central 
Article 

Google Scholar 
19.Saleh, N. & Chittka, L. Traplining in bumblebees (Bombus impatiens): a foraging strategy’s ontogeny and the importance of spatial reference memory in short-range foraging. Oecologia 151, 719–730 (2007).ADS 
PubMed 
Article 
PubMed Central 

Google Scholar 
20.Woodgate, J. L., Makinson, J. C., Lim, K. S., Reynolds, A. M. & Chittka, L. Continuous radar tracking illustrates the development of multi-destination routes of bumblebees. Sci. Rep. 7, 1–15 (2017).CAS 
Article 

Google Scholar 
21.Lihoreau, M., Chittka, L. & Raine, N. E. Trade-off between travel distance and prioritization of high-reward sites in traplining bumblebees. Funct. Ecol. 25, 1284–1292 (2011).PubMed 
PubMed Central 
Article 

Google Scholar 
22.Lihoreau, M., Chittka, L. & Raine, N. E. Travel optimization by foraging bumblebees through readjustments of traplines after discovery of new feeding locations. Am. Nat. 176, 744–757 (2010).PubMed 
Article 
PubMed Central 

Google Scholar 
23.Lihoreau, M., Chittka, L., Le Comber, S. C. & Raine, N. E. Bees do not use nearest-neighbour rules for optimization of multi-location routes. Biol. Lett. 8, 13–16 (2012).PubMed 
Article 
PubMed Central 

Google Scholar 
24.Minahan, D. F. & Brunet, J. Strong interspecific differences in foraging activity observed between honey bees and bumble bees using miniaturized radio frequency identification (RFID). Front. Ecol. Evol. 6, 156 (2018).Article 

Google Scholar 
25.Brunet, J., Zhao, Y. & Clayton, M. K. Linking the foraging behavior of three bee species to pollen dispersal and gene flow. PLoS ONE 14, e0212561 (2019).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
26.Reynolds, A. M., Lihoreau, M. & Chittka, L. A simple iterative model accurately captures complex trapline formation by bumblebees across spatial scales and flower arrangements. PLoS Comput. Biol. 9, e1002938 (2013).ADS 
MathSciNet 
CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
27.Marschall, E. A., Chesson, P. L. & Stein, R. A. Foraging in a patchy environment: prey-encounter rate and residence time distributions. Anim. Behav. 37, 444–454 (1989).Article 

Google Scholar 
28.Pyke, G. H. Optimal foraging theory : a critical review. Ann. Rev. Ecol. Syst. 15, 523–575 (1984).Article 

Google Scholar 
29.Rands, S. A. Landscape fragmentation and pollinator movement within agricultural environments: a modelling framework for exploring foraging and movement ecology. PeerJ 2, e269 (2014).PubMed 
PubMed Central 
Article 

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

Google Scholar 
31.Brunet, J. A conceptual framework that links pollinator foraging behavior to gene flow. In Proceedings for the 2018 Winter Seed Conference 63–67 (2018).32.Macarthur, R. H. & Pianka, E. R. On optimal use of a patchy environment. Am. Nat. 100, 603–609 (1966).Article 

Google Scholar 
33.Pyke, G. H. Understanding movements of organisms: it’s time to abandon the Lévy foraging hypothesis. Methods Ecol. Evol. 6, 1–16 (2015).Article 

Google Scholar 
34.Heinrich, B. ‘Majoring’ and ‘minoring’ by foraging bumblebees, Bombus vagans: an experimental analysis. Ecology 60, 245–255 (1979).Article 

Google Scholar 
35.Somme, L. et al. Pollen and nectar quality drive the major and minor floral choices of bumble bees. Apidologie 46, 92–106 (2015).ADS 
Article 

Google Scholar 
36.Levey, D. J., Bolker, B. M., Tewksbury, J. J., Sargent, S. & Haddad, N. M. Effects of landscape corridors on seed dispersal by birds. Science 309, 146–148 (2005).ADS 
CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
37.Levey, D. J., Tewksbury, J. J. & Bolker, B. M. Modelling long-distance seed dispersal in heterogeneous landscapes. J. Ecol. 96, 599–608 (2008).Article 

Google Scholar 
38.Pasquet, R. S. et al. Long-distance pollen flow assessment through evaluation of pollinator foraging range suggests transgene escape distances. Proc. Natl. Acad. Sci. U. S. A. 105, 13456–13461 (2008).ADS 
CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
39.Smith, K. & Spangenberg, G. Considerations for managing agricultural co-existence between transgenic and non-transgenic cultivars of outcrossing perennial forage plants in dairy pastures. Agronomy 6, 59–68 (2016).Article 

Google Scholar 
40.Ellstrand, N. C. et al. Introgression of crop alleles into wild or weedy populations. Annu. Rev. Ecol. Evol. Syst. 44, 325–345 (2013).Article 

Google Scholar 
41.Gupta, R. M. & Musunuru, K. Expanding the genetic editing tool kit: ZFNs, TALENs, and CRISPR-Cas9. J. Clin. Invest. 124, 4154–4161 (2014).PubMed 
PubMed Central 
Article 

Google Scholar 
42.Esch, H. E. & Burns, J. E. Distance estimation by foraging honeybees. J. Exp. Biol. 199, 155–162 (1996).CAS 
PubMed 
PubMed Central 

Google Scholar 
43.Srinivasan, M. V., Zhang, S., Altwein, M. & Tautz, J. Honeybee navigation: nature and calibration of the ‘odometer’. Science (80-.) 287, 851–853 (2000).ADS 
CAS 
Article 

Google Scholar 
44.Collett, M. & Collett, T. S. How do insects use path integration for their navigation?. Biol. Cybern. 83, 245–259 (2000).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
45.Collett, M., Chittka, L. & Collett, T. S. Spatial memory in insect navigation. Curr. Biol. 23, R789–R800 (2013).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
46.Chittka, L., Geiger, K. & Kunze, J. The influences of landmarks on distance estimation of honey bees. Anim. Behav. 50, 23–31 (1995).Article 

Google Scholar 
47.Srinivasan, M. V., Lehrer, M. & Horridge, G. A. Visual figure-ground discrimination in the honeybee: the role of motion parallax at boundaries. Proc. R. Soc. B Biol. Sci. 238, 331–350 (1990).ADS 

Google Scholar 
48.Lehrer, M. Looking all around: honeybees use different cues in different eye regions. J. Exp. Biol. 201, 3275–3292 (1998).PubMed 
PubMed Central 

Google Scholar 
49.Goulson, D. Foraging strategies of insects for gathering nectar and pollen, and implications for plant ecology and evolution. Perspect. Plant Ecol. Evol. Syst. 2, 185–209 (1999).Article 

Google Scholar 
50.Ohashi, K., Thomson, J. D. & D’Souza, D. Trapline foraging by bumble bees: IV. Optimization of route geometry in the absence of competition. Behav. Ecol. 18, 1–11 (2007).Article 

Google Scholar 
51.Comba, L. Patch use by bumblebees (hymenoptera apidae): temperature, wind, flower density and traplining. Ethol. Ecol. Evol. 11, 243–264 (1999).Article 

Google Scholar 
52.Ohashi, K., Leslie, A. & Thomson, J. D. Trapline foraging by bumble bees: V. Effects of experience and priority on competitive performance. Behav. Ecol. 19, 936–948 (2008).Article 

Google Scholar 
53.Klein, S., Pasquaretta, C., Barron, A. B., Devaud, J. M. & Lihoreau, M. Inter-individual variability in the foraging behaviour of traplining bumblebees. Sci. Rep. 7, 1–12 (2017).Article 
CAS 

Google Scholar 
54.Chittka, L. Bee cognition. Curr. Biol. 27, R1049–R1053 (2017).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
55.Ohashi, K. & Yahara, T. Visit larger displays but probe proportionally fewer flowers: counterintuitive behaviour of nectar-collecting bumble bees achieves an ideal free distribution. Funct. Ecol. 16, 492–503 (2002).Article 

Google Scholar 
56.Brunet, J. & Stewart, C. M. Impact of bee species and plant density on alfalfa pollination and potential for gene flow. Psyche A J. Entomol. https://doi.org/10.1155/2010/201858 (2010).Article 

Google Scholar 
57.R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2019).
Google Scholar 
58.Weisberg, S. Applied Linear Regression (Wiley, 2013). https://doi.org/10.2307/3150981.Book 
MATH 

Google Scholar  More

1.Kellar, N. M. et al. Blubber testosterone: a potential marker of male reproductive status in short-beaked common dolphins. Mar. Mamm. Sci. 25(3), 507–522 (2009).CAS 
Article 

Google Scholar 
2.Atkinson, S. & Yoshioka, M. Endocrinology of reproduction. In Reproductive biology and phylogeny of Cetacea. Whales, dolphins and porpoises (ed. Miller, D. L.) 171–192 (Science Publishers, 2007).
Google Scholar 
3.Cates, K. A. et al. Testosterone trends within and across seasons in male humpback whales (Megaptera novaeangliae) from Hawaii and Alaska. Gen. Comp. Endocrinol. 279, 164–173 (2019).CAS 
PubMed 
Article 

Google Scholar 
4.McKenna, T. J. et al. 2 A critical review of the origin and control of adrenal androgens. Baillieres Clin. Obstet. Gynaecol. 11(2), 229–248 (1997).CAS 
PubMed 
Article 

Google Scholar 
5.Sharpe, R. et al. Testosterone and Spermatogenesis identification of stage-specific, androgen-regulated proteins secreted by adult rat seminiferous tubules. J. Androl. 13, 172–184 (1992).CAS 
PubMed 

Google Scholar 
6.Kita, S., Yoshioka, M. & Kashiwagi, M. Relationship between sexual maturity and serum and testis testosterone concentrations in short-finned pilot whales Globicephala macrorhynchus. Fish. Sci. 65(6), 878–883 (1999).CAS 
Article 

Google Scholar 
7.Schroeder, J. P. & Keller, K. V. Seasonality of serum testosterone levels and sperm density in Tursiops truncatus. J. Exp. Zool. 249(3), 316–321 (1989).CAS 
PubMed 
Article 

Google Scholar 
8.Robeck, T. R. et al. Reproduction, growth and development in captive beluga (Delphinapterus leucas). Zoo Biol. 24(1), 29–49 (2005).Article 

Google Scholar 
9.Wells, R. Reproductive behavior and hormonal correlates in Hawaiian spinner dolphins, Stenella longirostris. In Reproduction in Whales, Dolphins, and Porpoises (eds Perrin, W. F. et al.) 465–472 (Reports of the International Whaling Commission, 1984).
Google Scholar 
10.Mogoe, T. et al. Functional reduction of the southern minke whale (Balaenoptera acutorostrata) testis during the feeding season. Mar. Mamm. Sci. 16(3), 559–569 (2000).Article 

Google Scholar 
11.Kjeld, M. et al. Changes in blood testosterone and progesterone concentrations of the North Atlantic minke whale (Balaenoptera acutorostrata) during the feeding season. Can. J. Fish. Aquat. Sci. 61(2), 230–237 (2004).CAS 
Article 

Google Scholar 
12.Temte, J. L. Use of serum progesterone and testosterone to estimate sexual maturity in Dall’s porpoise Phocoenoides dalli. Fish. Bull. 89(1), 161–166 (1991).
Google Scholar 
13.Robeck, T. R. et al. Reproduction, growth and development in captive beluga (Delphinapterus leucas). Zoo Biol. Publ. Affil. Am. Zoo Aquar. Assoc. 24(1), 29–49 (2005).
Google Scholar 
14.Kirby, V. L. Endocrinology of marine mammals. In Handbook of Marine Mammal Medicine: Health (ed. Dierauf, L. A.) 303–351 (Disease and Rehabilitation CRC Press Inc, 1990).
Google Scholar 
15.Desportes, G., Saboureau, M. & Lacroix, A. Growth-related changes in testicular mass and plasma testosterone concentrations in long-finned pilot whales, Globicephala melas. J. Reprod. Fertil. 102(1), 237–244 (1994).CAS 
PubMed 
Article 

Google Scholar 
16.Kjeld, J., Sigurjonsson, J. & Arnason, A. Sex hormone concentrations in blood serum from the North Atlantic fin whale (Balaenoptera physalus). J. Endocrinol. 134(3), 405–413 (1992).CAS 
PubMed 
Article 

Google Scholar 
17.Boggs, A. S. et al. Remote blubber sampling paired with liquid chromatography tandem mass spectrometry for steroidal endocrinology in free-ranging bottlenose dolphins (Tursiops truncatus). Gen. Comp. Endocrinol. 281, 164–172 (2019).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
18.Weller, D. W. et al. Behavioral responses of bottlenose dolphins to remote biopsy sampling and observations of surgical biopsy wound healing. Aquat. Mamm. 23(1), 49–58 (1997).19.Krahn, M. M. et al. Stratification of lipids, fatty acids and organochlorine contaminants in blubber of white whales and killer whales. J. Cetac. Res. Manage. 6(2), 175–189 (2004).
Google Scholar 
20.Marsili, L. et al. Skin biopsies for cell cultures from Mediterranean free-ranging cetaceans. Mar. Environ. Res. 50(1–5), 523–526 (2000).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
21.Hobbs, K. E. et al. PCBs and organochlorine pesticides in blubber biopsies from free-ranging St. Lawrence River Estuary beluga whales (Delphinapterus leucas), 1994–1998. Environ. Pollut. 122(2), 291–302 (2003).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
22.Kellar, N. M. et al. Determining pregnancy from blubber in three species of delphinids. Mar. Mamm. Sci. 22(1), 1–16 (2006).Article 

Google Scholar 
23.Mingramm, F. et al. Evaluation of respiratory vapour and blubber samples for use in endocrine assessments of bottlenose dolphins (Tursiops spp.). Gen. Comp. Endocrinol. 274, 37–49 (2019).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
24.Galligan, T. M. et al. Blubber steroid hormone profiles as indicators of physiological state in free-ranging common bottlenose dolphins (Tursiops truncatus). Comp. Biochem. Physiol. A Mol. Integr. Physiol. 239, 110583 (2020).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
25.Dierauf, L. & Gulland, F. M. CRC Handbook of Marine Mammal Medicine: Health, Disease, and Rehabilitation (CRC Press, 2001).Book 

Google Scholar 
26.Champagne, C. D. et al. Comprehensive endocrine response to acute stress in the bottlenose dolphin from serum, blubber, and feces. Gen. Comp. Endocrinol. 266, 178–193 (2018).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
27.Kellar, N. M. et al. Variation of bowhead whale progesterone concentrations across demographic groups and sample matrices. Endang. Species Res. 22(1), 61–72 (2013).Article 

Google Scholar 
28.Richard, J. T. et al. Testosterone and progesterone concentrations in blow samples are biologically relevant in belugas (Delphinapterus leucas). Gen. Comp. Endocrinol. 246, 183–193 (2017).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
29.Hunt, K. E. et al. Multi-year patterns in testosterone, cortisol and corticosterone in baleen from adult males of three whale species. Conserv. Physiol. 6(1), coy049 (2018).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
30.Wells, R. S. Dolphin social complexity: lessons from long-term study and life history. In Animal Social Complexity: Intelligence, Culture, and Individualized Societies (eds de Waal, F. B. M. & Tyack, P. L.) 32–56 (Harvard University Press, 2003).
Google Scholar 
31.Brook, F. et al. Ultrasonographic imaging of the testis and epididymis of the bottlenose dolphin, Tursiops truncatus aduncas. J. Reprod. Fertil. 119(2), 233–240 (2000).CAS 
PubMed 
Article 

Google Scholar 
32.Wells, R. S. Social structure and life history of bottlenose dolphins near Sarasota Bay, Florida: Insights from four decades and five generations. In Primates and Cetaceans: Field Research and Conservation of Complex Mammalian Societies, Primatology Monographs (eds Yamagiwa, J. & Karczmarski, L.) 149–172 (Springer, 2014).
Google Scholar 
33.Barratclough, A. et al. Health assessments of common bottlenose dolphins (Tursiops truncatus): past, present, and potential conservation applications. Front. Vet. Sci. 6, 444 (2019).PubMed 
PubMed Central 
Article 

Google Scholar 
34.Champagne, C. D. et al. Blubber cortisol qualitatively reflects circulating cortisol concentrations in bottlenose dolphins. Mar. Mamm. Sci. 33(1), 134–153 (2017).CAS 
Article 

Google Scholar 
35.Norman, A. W. & Litwack, G. Hormones (Academic Press, 1997).
Google Scholar 
36.Urian, K. et al. Seasonality of reproduction in bottlenose dolphins, Tursiops truncatus. J. Mammal. 77(2), 394–403 (1996).Article 

Google Scholar 
37.Wells, R. Reproduction in wild bottlenose dolphins: overview of patterns observed during a long-term study. in Bottlenose Dolphins Reproduction Workshop. AZA marine mammal taxon advisory group. 2000. Silver Springs, MD.38.Read, A. et al. Patterns of growth in wild bottlenose dolphins, Tursiops truncatus. J. Zool. 231(1), 107–123 (1993).Article 

Google Scholar 
39.Trego, M. L. et al. Comprehensive screening links halogenated organic compounds with testosterone levels in male Delphinus delphis from the Southern California bight. Environ. Sci. Technol. 52(5), 3101–3109 (2018).ADS 
CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
40.Kannan, K. et al. Toxicity reference values for the toxic effects of polychlorinated biphenyls to aquatic mammals. Hum. Ecol. Risk Assess. 6(1), 181–201 (2000).CAS 
Article 

Google Scholar 
41.Jepson, P. D. et al. Relationships between polychlorinated biphenyls and health status in harbor porpoises (Phocoena phocoena) stranded in the United Kingdom. Environ. Toxicol. Chem. Int. J. 24(1), 238–248 (2005).CAS 
Article 

Google Scholar 
42.Minter, L. & DeLiberto, T. Seasonal variation in serum testosterone, testicular volume, and semen characteristics in the coyote (Canis latrans). Theriogenology 69(8), 946–952 (2008).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
43.Preston, B. T. et al. Testes size, testosterone production and reproductive behaviour in a natural mammalian mating system. J. Anim. Ecol. 81(1), 296–305 (2012).PubMed 
Article 
PubMed Central 

Google Scholar 
44.Desportes, G., Saboureau, M. & Lacroix, A. Growth-related changes in testicular mass and plasma testosterone concentrations in long-finned pilot whales Globicephala melas. Reproduction 102(1), 237–244 (1994).CAS 
Article 

Google Scholar 
45.Ryan, C. et al. Lipid content of blubber biopsies is not representative of blubber in situ for fin whales (Balaenoptera physalus). Mar. Mamm. Sci. 29(3), 542–547 (2013).CAS 
Article 

Google Scholar 
46.Wells, R. S. et al. Bottlenose dolphins as marine ecosystem sentinels: developing a health monitoring system. EcoHealth 1(3), 246–254 (2004).Article 

Google Scholar 
47.Wells, R. S. et al. Integrating life-history and reproductive success data to examine potential relationships with organochlorine compounds for bottlenose dolphins (Tursiops truncatus) in Sarasota Bay Florida. Sci. Total Environ. 349(1–3), 106–119 (2005).ADS 
CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
48.Yordy, J. E. et al. Partitioning of persistent organic pollutants between blubber and blood of wild bottlenose dolphins: implications for biomonitoring and health. Environ. Sci. Technol. 44(12), 4789–4795 (2010).ADS 
CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
49.Kellar, N. M. et al. Low reproductive success rates of common bottlenose dolphins Tursiops truncatus in the northern Gulf of Mexico following the Deepwater Horizon disaster (2010–2015). Endang. Species Res. 33, 143–158 (2017).Article 

Google Scholar 
50.Kellar, N. M. et al. Blubber cortisol: a potential tool for assessing stress response in free-ranging dolphins without effects due to sampling. PLoS ONE 10(2), e0115257 (2015).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
51.Harrison, R. & Ridgway, S. Gonadal activity in some bottlenose dolphins (Tursiops truncatus). J. Zool. 165(3), 355–366 (1971).Article 

Google Scholar 
52.Wells, R. S. & Scott, M. D. Bottlenose Dolphin: common bottlenose dolphin: Tursiops truncates. In Encyclopedia of Marine Mammals 3rd edn (eds Würsig, B. et al.) 118–125 (Academic Press/Elsevier, 2009).
Google Scholar 
53.R Core Team. R: A Language and Environment for Statistical Computing. 2020, R Foundation for Statistical Computing: 2020, Vienna, Austria. URL https://www.R-project.org/.54.Wells, R. S. & Scott, M. D. Common Bottlenose Dolphin: Tursiops truncates, in Encyclopedia of Marine Mammals 252 (Elsevier, 2009).
Google Scholar  More

Let us first describe the setting and introduce notation. The main object of analysis is a matrix A (a contingency table with (n_r) rows and (n_c) columns) that contains the counts of two variables. A common example from ecology is that (A_{ij}) contains some measure of abundance of species i (rows) in sampling site j (columns). The matrix A can also be a binary incidence matrix, containing either the presence (1) or absence (0) of species in sites. The matrix A can be interpreted as the bi-adjacency matrix of a bipartite network that connects species to sites. The network contains (n_r) nodes on one side (the species, given by the rows of A, indexed by i), and (n_c) nodes on the other side (the sites, given by the columns of A, indexed by j). In general, we will refer to the two sets of nodes as row nodes and column nodes, respectively. The degree of a row node i is defined by the row sum (r_i = sum _j A_{ij}), which gives the total abundance of species i in all sites. Likewise, the degree of a column node j is defined as the column sum (c_j = sum _i A_{ij}), which gives the total abundance of species in a site j. The degrees of the row and column nodes are given by the vectors (mathbf {r}= (r_1, r_2, dots , r_{n_r})^T) and (mathbf {c}= (c_1,c_2,dots ,c_{n_c})^T). We further define two square matrices, (D_r) ((n_r times n_r)) and (D_c) ((n_c times n_c)) as the diagonal matrices that have (mathbf {r}) and (mathbf {c}) on the diagonal, respectively. The sum (n = sum _{ij} A_{ij}) gives the total number of occurrences in the table (in the case of a species-site example, the total abundance of species).CA as canonical correlation analysisOne of the first derivations of CA was obtained by applying canonical correlation analysis to categorical variables12,19,22. Here we follow the derivation in Ref.1 (Chapter 9), where CA is derived as an application of canonical correlation analysis applied to a bipartite network, and to which we refer for further details. For ease of explanation, we will assume the network is defined by a binary presence-absence matrix (i.e. the network is unweighted), but the result generalizes to any contingency table (i.e. weighted bipartite networks).The aim is to assign a ‘score’ to each node in the network, under the assumption that row and column nodes with similar scores connect to each other. Hence, connected nodes get assigned similar scores, and the scores can be thought of as a latent variable that drive the formation of links in the network. In ecology, such latent variables are referred to as gradients2,3. Considering a bipartite network describing the occurrence of species in a set of sites, for example, the resulting scores may reflect some variable determining why species locate in specific sites, such as the temperature preference of a species and the temperature at a site. In practice, the interpretation of a gradient resulting from application of CA can be verified by correlating it with known environmental variables (e.g. data on the temperature of each site).Mathematically, such gradients can be inferred from the edges of the bipartite network. Recall that for a presence-absence matrix, the total number of edges in the bipartite network is given by (n = sum _{ij} A_{ij}). Let us construct a vector (mathbf {y}_r) of length n that contains, for each edge, the scores of the row node it connects to, and a vector (mathbf {y}_c) of length n that contains, again for each edge, the score of the column node it connects to. Given the assumption that edges connect row nodes and column nodes with similar scores, the node scores can be found by maximizing the correlation between (mathbf {y}_r) and (mathbf {y}_c), so that the row- and column scores for each edge are as similar as possible. Denoting the vector of length (n_r) containing the row scores by (mathbf {v}) and the vector of length (n_c) containing the column scores by (mathbf {u}), this leads to the optimization problem$$begin{aligned} max _{mathbf {v}, mathbf {u}} mathrm {corr}(mathbf {y}_r,mathbf {y}_c). end{aligned}$$
(1)
In order to obtain standardized scores, the constraints that (mathbf {y}_r) and (mathbf {y}_c) have zero mean and unit variance need to be added. Solving this problem using Lagrangian optimization, the solution is given by$$begin{aligned} D_r^{-1}A D_c^{-1} A^T mathbf {v}&= lambda mathbf {v}nonumber \ D_c^{-1}A^T D_r^{-1} A mathbf {u}&= lambda mathbf {u}. end{aligned}$$
(2)
The score vectors (mathbf {v}) and (mathbf {u}) can thus be found by solving an eigenvector problem. Following Ref.1 , they are subject to the constraint that (|mathbf {v}|=|mathbf {u}|=1). The general interpretation of the elements of (mathbf {v}) and (mathbf {u}) is as follows. Each row node (of A) is represented in (mathbf {v}), and each column node is represented in (mathbf {u}). The smaller the difference between the values of two row (column) nodes in (mathbf {v}) ((mathbf {u})), the more similar these nodes are. The similarity among row nodes that is reflected in (mathbf {v}) vectors arises because they are connected to a similar set of column nodes in the original bipartite network (vice versa for similarities in (mathbf {u})). In literature, this is referred to as row nodes being similar because of their similar ‘profile’11,23, and reciprocal averaging defines exactly how the scores are calculated in terms of the profiles and reduces to the same set of equations15. Both matrices on the left-hand side of Eq. (2) are row-stochastic and positive definite, and have identical eigenvalues that are real and take values between 0 and 1. Assuming that we have a connected network, sorting the eigenvalues in decreasing order leads to (1=lambda _1 > lambda _2 dots ge 0).It can be shown that the correlation between (mathbf {y}_r) and (mathbf {y}_c) for a given set of eigenvectors (mathbf {v}) and (mathbf {u}) is given by their corresponding eigenvalue, so that (lambda = mathrm {corr}^2(mathbf {y}_{r},mathbf {y}_{c})). Note that the correlations between the row and column vectors can be negative, meaning that merely the absolute value of the correlation between (y_r) and (y_c) is related to the (square root) of the eigenvalues. Iterative approaches to extract potential negative correlations exist in literature24. The node scores leading to the highest correlation are thus given by the eigenvectors associated with the largest eigenvalue. However, the eigenvectors corresponding to (lambda _1) have all constant values and represent the trivial solution in which all row nodes and all column nodes have equal scores (leading to a perfect correlation). This trivial solution does not satisfy the condition that the scores have to be centered, and thus it must be rejected. The solution to Eq. (1) is thus given by the eigenvectors (mathbf {v}_2) and (mathbf {u}_2), corresponding to the second largest eigenvalue (lambda _2), which corresponds to the square root of the (maximized) correlation. We notice here that this derivation leads to analogous results than observed in classical derivations of CA, where the matrix A is centered both with respect to the rows and to the columns.The second eigenvectors (mathbf {v}_2) and (mathbf {u}_2) hold the unique scores such that row- and column nodes with similar scores connect to each other. The second eigenvalue (lambda _2) indicates to what extent the row- and column scores can be ‘matched’, where high values (close to 1) indicate a high association between the inferred scores (the gradient) and the structure of the network.The higher order eigenvectors in Eq. (2) and their eigenvalues are solutions to Eq. (1) with the additional constraint that (mathbf {y}_r) and (mathbf {y}_c) are orthogonal to the other solutions. The vectors (mathbf {v}_3) and (mathbf {u}_3), for example, may represent other variables that drive the formation of links (e.g. precipitation, primary productivity, etc.) on top of the gradients described by (mathbf {v}_2) and (mathbf {u}_2). We note that, differently from notation in some CA literature, we here denote the k-th non-trivial eigenvector with the subscript k+1.CA as a clustering algorithmA completely different approach shows that the eigenvectors (mathbf {v}_2) and (mathbf {u}_2) (i.e. the second eigenvectors in Eq. (2)) can also be interpreted as cluster labels, obtained when identifying clusters in the network of similarities that is derived from the bipartite network.A similarity network can be constructed from a bipartite network by ‘projecting’ the bipartite network onto one of its layers (either the row nodes or the column nodes) through stochastic complementation18. Projecting the bipartite network defined by A onto its row layer leads to the (n_r times n_r) similarity matrix (S_r = A D_c^{-1} A^T). The entries of (S_r) represent pairwise similarities between row nodes of A, based on how many links they share with the same column node, weighted for the degree of each column node. Similarly, the (n_c times n_c) similarity matrix (S_c = A^T D_r^{-1} A) defines the pairwise similarities between the column nodes of A.Identifying clusters in the similarity network can be done by minimizing the so-called ‘normalized cut’20. The normalized cut assigns, for a given partition of a network into K clusters, a score that represents the strength of the connections between the clusters for that partition. A partition can be described by assigning a discrete cluster label to each node. Hence, minimizing the normalized cut is equivalent to assigning a cluster label to each node in the network in such a way that the clusters are minimally connected. Finding the discrete cluster labels that minimize the normalized cut in large networks is in general not possible20. However, a solution of a related problem can be obtained when the cluster labels are allowed to take continuous values as opposed to discrete values. Solutions of this ‘relaxed’ problem can be interpreted as continuous approximations of the discrete cluster labels.Minimizing the normalized cut in (S_r) leads to the generalized eigensystem20$$begin{aligned} (D_r – S_r) mathbf {v}= tilde{lambda } D_r mathbf {v}, end{aligned}$$
(3)
where the entries of the generalized eigenvector (mathbf {v}_2) corresponding to the second smallest eigenvalue (tilde{lambda }_2) of Eq. (3) hold the approximate cluster labels of the optimal partition into two clusters. It is easily shown that generalized eigenvectors in Eq. (3) are exactly the eigenvectors of Eq. (2), where the eigenvalues are related by (tilde{lambda }_k = 1 – lambda _k), where (k=1,2,dots ,n_r) (see “Suppl. Material A”).The matrix (L_r = D_r-S_r) is known as the Laplacian matrix of the similarity network defined by (S_r), and is well known in spectral graph theory25. The number of eigenvalues of (L_r) for which (tilde{lambda } = 0) (or equivalently (lambda = 1) in Eq. (2)) denotes the number of disconnected clusters in the network. Each of these ’trivial’ eigenvalues has a corresponding generalized eigenvector that has constant values for the nodes in a particular cluster, indicating cluster membership.The situation changes when the clusters are weakly connected. The optimal solution for partitioning the similarity network into two clusters is given by the eigenvector (mathbf {v}_2) associated to eigenvalue (lambda _2). Although continuous, the entries of (mathbf {v}_2) can be interpreted as approximations to cluster labels, which indicate for each row node to which cluster it belongs. In other words, nodes with high values in this eigenvector (i.e., high scores) belong to one cluster, and nodes with low scores to the other. A discrete partition can then be obtained from the approximate (continuous) cluster labels by discretizing them, for example by assigning all negative values to one cluster and all positive values to the other26. The corresponding eigenvalue (lambda _2) represents the quality of the partitioning, as determined by the normalized cut criterion. High values are indicative of a network that can be well partitioned into two clusters (two totally disconnected clusters would yield eigenvalues (lambda _1 = lambda _2=1)), whereas lower values correspond to a network that is less easily grouped into two clusters (i.e. the resulting clusters are more interconnected).Finding a partitioning into multiple, say K, clusters is more involved, where (Kle n_r) (or (Kle n_c) if working with column variables). Minimizing the normalized cut for K clusters yields a trace minimization problem of which the relaxed solution is given by the first K eigenvectors in Eq. (2)27. Because the first eigenvector in Eq. (2) is trivial, in practice we only require (K-1) eigenvectors (i.e., the 2nd, 3rd, … up to the Kth). The discrete cluster labels can then be obtained, for example, by running a k-Means algorithm on the matrix consisting of those (K-1) eigenvectors, a technique that is also known as spectral clustering28,29. How well the network can be partitioned into K clusters is given by the average value of the first K eigenvalues, i.e. (frac{1}{K} sum _{k=1}^K lambda _k)27.The clustering approach thus brings an alternative interpretation to CA results. A key observation is that the eigenvalues and eigenvectors in Eq. (2) are directly related to the generalized eigenvectors of the Laplacian of the similarity matrix (S_r), and thus hold information on the structure of the similarity network. The entries of the second eigenvector (mathbf {v}_2) can be interpreted as the approximate cluster labels of a two-way partitioning of the similarity network defined by (S_r). Although at first sight the interpretation of CA scores as cluster labels may seem different from the interpretation as a latent variable described above in “CA as canonical correlation analysis”, note that cluster labels can be seen as latent variables, albeit discrete rather than continuous.CA as a graph embedding techniqueA third interpretation of the eigenvectors and eigenvalues in Eq. 2 arises from a so-called graph embedding of the similarity matrix (S_r) (or (S_c)). Graph embeddings provide a way to obtain a low-dimensional representation of a high-dimensional network, that are used for example for graph drawing. A graph embedding represents the nodes of a graph as node vectors in a space, such that nodes that are ‘close’ in the network are also close in terms of their distance in the embedding. A key feature of these embeddings is that their dimensionality can be reduced in order to obtain a low-dimensional representation of the data, while retaining its most important structural properties (see Ref.1, chapter 10 for an overview of graph embedding techniques).As noted by several authors, CA is equivalent to graph embedding in the case of a similarity matrix obtained through stochastic complementation. For example, computing a 1-step diffusion map of the similarity matrix (S_r) leads exactly to the eigenvectors of Eq. (2)18,30. Also, the graph embedding using the Laplacian eigenmap has been shown to be equivalent to graph partitioning using the normalized cut, which is in turn equivalent to CA31. CA-specific type of embedding is based on the chi-square statistics and it is thus Euclidean.Embedding the similarity network (S_r) in a ((K-1))-dimensional space yields an ‘embedding matrix’ (X_r in mathbb {R}^{n_r times K-1}) (known in CA-related literature as ’principal coordinates’). Each row of (X_r) represents a node of (S_r) as a ‘node vector’ in the embedding. The rows of (X_r) can be seen as components of ((K-1))-dimensional basis vectors that span the embedding, and are identical to what is referred to as the ‘axes’ in CA. Every entry (X_{i,k}) represents the coordinate of row node i on the k’th basis vector, and can be seen as the ‘score’ of i on the k’th CA axis. An embedding matrix of (S_r) can defined as (X_r = [sqrt{lambda _2} mathbf {v}_2, dots , sqrt{lambda _K} mathbf {v}_{K}]), where the vectors (mathbf {v}_k) are the eigenvectors defined in (2), and each of them is weighted by the square root of their corresponding eigenvalue. We will refer to columns of the embedding matrix as ‘CA-axes’, given by (mathbf {x}_k = sqrt{lambda _k}mathbf {v}_k) (with (mathbf {x}_2) being the ’first CA axis’, and so on).The axes are constructed in such a way that they capture the largest amount of ‘variation’ or ‘inertia’ in the data, which is given by their corresponding eigenvalue11. The sum of all the eigenvalues gives the total variation in the data (in CA, this is referred to as the total inertia). CA decomposes the total variation in such a way that the first axis captures a maximal part of the variation, the second a maximal part of the remaining variation, and so on. A low-dimensional embedding that preserves the maximal amount of variation can thus be obtained by discarding the eigenvectors corresponding to smaller eigenvalues. The ‘quality’ of the embedding can then be expressed as the share of the total variation that is preserved in the embedding.A typical way of presenting CA results is by showing the first two coordinates of each row (or column) node, i.e. plotting (mathbf {x}_2) against (mathbf {x}_3), which is usually referred to as a ’correspondence plot’. Since the first two axes capture a maximal amount of inertia, such a plot is in a way the optimal two-dimensional representation of the data that captures the relations between the rows (or columns) of A. The distances between points in the correspondence plot approximate the similarities between nodes. How well the correspondence plot represents the similarities is given by the percentage of variation explained by the first two axes.Each axis can be interpreted as a latent variable that account for part of the total variation in the data. Since the axes in the embedding are given by a scaled version of the eigenvectors discussed above in “CA as canonical correlation analysis”, the interpretation of the eigenvalues as the amount of variation explained is complementary to the interpretation as the correlation between row and column scores which we also introduced above in “CA as canonical correlation analysis”. Furthermore, the axes spanning the K-dimensional embedding are exactly the generalized eigenvectors that follow from minimizing the normalized cut for K clusters31. Indeed, when there are clear clusters in the similarity network, they will show up in the embedding space as separate groups of points.Summarizing, we find three interpretations of CA axes and their corresponding eigenvalues: as latent variables that drive the formation of links in the bipartite network, as approximate clusters labels of a bi-partition of the similarity network, and as coordinates of an embedding of the similarity network. The different derivations of CA and their interpretations are summarized in Table 1.Table 1 Different interpretations of the eigenvectors and eigenvalues resulting from CA.Full size table More

1.Bradford, M. A. et al. Climate fails to predict wood decomposition at regional scales. Nat. Clim. Chan. 4, 625–630 (2014).CAS 
Article 
ADS 

Google Scholar 
2.Crowther, T. W. et al. The global soil community and its influence on biogeochemistry. Science 365, eaav0550 (2019).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
3.Lustenhouwer, N. et al. A trait-based understanding of wood decomposition by fungi. Proc. Nat. Acad. Sci. USA 117, 11551–11558 (2020).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
4.Tilman, D. et al. Diversity and productivity in a long-term grassland experiment. Science 294, 843–845 (2001).CAS 
PubMed 
Article 
ADS 
PubMed Central 

Google Scholar 
5.Ammer, C. Diversity and forest productivity in a changing climate. New Phytol. 221, 50–66 (2019).PubMed 
Article 
PubMed Central 

Google Scholar 
6.Dickie, I. A., Fukami, T., Wilkie, J. P., Allen, R. B. & Buchanan, P. K. Do assembly history effects attenuate from species to ecosystem properties? A field test with wood-inhabiting fungi. Ecol. Lett. 15, 133–141 (2012).PubMed 
Article 
PubMed Central 

Google Scholar 
7.van der Wal, A., Ottosson, E. & de Boer, W. Neglected role of fungal community composition in explaining variation in wood decay rates. Ecology 96, 124–133 (2015).PubMed 
Article 
PubMed Central 

Google Scholar 
8.Hoppe, B. et al. Linking molecular deadwood-inhabiting fungal diversity and community dynamics to ecosystem functions and processes in Central European forests. Fung. Div. 77, 367–379 (2016).Article 

Google Scholar 
9.Purahong, W. et al. Determinants of deadwood-inhabiting fungal communities in temperate forests: Molecular evidence from a large scale deadwood decomposition experiment. Front. Microbiol. 9, Article 2120 (2018).10.Skelton, J. et al. Relationships among wood-boring beetles, fungi, and the decomposition of forest biomass. Mol. Ecol. 28, 4971–4986 (2019).CAS 
PubMed 
Article 

Google Scholar 
11.Toljander, Y. K., Lindahl, B. D., Holmer, L. & Hogberg, N. O. S. Environmental fluctuations facilitate species co-existence and increase decomposition in communities of wood decay fungi. Oecologia 148, 625–631 (2006).PubMed 
Article 
ADS 

Google Scholar 
12.Fukami, T. et al. Assembly history dictates ecosystem functioning: evidence from wood decomposer communities. Ecol. Lett. 13, 675–684 (2010).PubMed 
Article 

Google Scholar 
13.Boddy, L. Fungal community ecology and wood decomposition process in angiosperms: From standing tree to complete decay of coarse woody debris. Ecol. Bull. 49, 43–56 (2001).
Google Scholar 
14.Boddy, L. & Heilmann-Clausen, J. Basidiomycete community development in temperate angiosperm wood. In Ecology of saprotrpophic basidiomycetes. (Eds. Boddy, L., Frankland, J.C., & van West, P.) 211–237 (Academic Press, 2008).15.Parfitt, D., Hunt, J., Dockrell, D., Rogers, H. J. & Boddy, L. Do all trees carry the seed of their own destruction? PCR reveals numerous wood decay fungi latently present in sapwood of a wide range of angiosperm trees. Fung. Ecol. 3, 338–346 (2010).Article 

Google Scholar 
16.Song, Z., Kennedy, P. G., Liew, F. J. & Schilling, J. S. Fungal endophytes as priority colonizers initiating wood decomposition. Func. Ecol. 31, 407–418 (2017).Article 

Google Scholar 
17.Cline, L. C., Schilling, J. S., Menke, J., Groenhof, E. & Kennedy, P. G. Ecological and functional effects of fungal endophytes on wood decomposition. Func. Ecol. 32, 181–191 (2018).Article 

Google Scholar 
18.Coates, D. & Rayner, A. D. M. Fungal population and community development in cut beech logs I. Establishment via the aerial cut surface. New Phytol. 101, 153–171 (1985).Article 

Google Scholar 
19.Fukasawa, Y., Osono, T. & Takeda, H. Beech log decomposition by wood-inhabiting fungi in a cool temperate forest floor: A quantitative analysis focused on the decay activity of a dominant basidiomycetes Omphalotus guepiniformis. Ecol. Res. 25, 959–966 (2010).Article 

Google Scholar 
20.Boddy, L. & Hiscox, J. Fungal ecology: principles and mechanisms of colonization and competition by saprotrophic fungi. Microbiol. Spec. 4, FUNK-0019-2016 (2016).
Google Scholar 
21.Rajala, T., Peltoniemi, M., Pennanen, T. & Makipaa, R. Fungal community dynamics in relation to substrate quality of decaying Norway spruce (Picea abies [L.] Karst.) logs in boreal forests. FEMS Microbiol. Ecol. 81, 494–505 (2012).CAS 
PubMed 
Article 

Google Scholar 
22.Rajala, T., Tuomivirta, T., Pennanen, T. & Mäkipää, R. Habitat models of wood-inhabiting fungi along a decay gradient of Norway spruce logs. Fung. Ecol. 18, 48–55 (2015).Article 

Google Scholar 
23.Rayner, A.D.M., & Boddy, L. Fungal decomposition of wood: Its biology and ecology. (Willey, 1988).24.Bunnell, F. L. & Houde, I. Down wood and biodiversity—Implications to forest practices. Environ. Rev. 18, 397–421 (2010).Article 

Google Scholar 
25.Wells, J. M. & Boddy, L. Interspecific carbon exchange and cost of interactions between basidiomycete mycelia in soil and wood. Func. Ecol. 16, 153–161 (2002).Article 

Google Scholar 
26.Hiscox, J. et al. Effects of pre-colonisation and temperature on interspecific fungal interactions in wood. Fung. Ecol. 21, 32–42 (2016).Article 

Google Scholar 
27.Fukasawa, Y., Osono, T. & Takeda, H. Wood decomposition abilities of diverse lignicolous fungi on nondecayed and decayed beech wood. Mycologia 103, 474–482 (2011).PubMed 
Article 

Google Scholar 
28.Valentin, L. et al. Loss of diversity in wood-inhabiting fungal communities affects decomposition activity in Norway spruce wood. Front. Microbiol. 5, Article 230 (2014).29.Maynard, D., Crowther, T. W. & Bradford, M. A. Fungal interactions reduce carbon use efficiency. Ecol. Lett. 20, 1034–1042 (2017).PubMed 
Article 

Google Scholar 
30.Woodward, S., & Boddy, L. Interactions between saprotrophic fungi. In Ecology of saprotrophic basidiomycetes (eds Boddy, L., Frankland, J.C., van West, P.) 125–141 (Academic Press, 2008).31.Gessner, M. O. et al. Diversity meets decomposition. Trends Ecol. Evol. 25, 372–380 (2010).PubMed 
Article 
PubMed Central 

Google Scholar 
32.Fukasawa, Y., Gilmartin, E. C., Savoury, M. & Boddy, L. Inoculum volume effects on competitive outcome and wood decay rate of brown- and white-rot basidiomycetes. Fung. Ecol. 45, 100938 (2020).Article 

Google Scholar 
33.O’Leary, J. et al. The whiff of decay: Linking volatile production and extracellular enzymes to outcomes of fungal interactions at different temperatures. Fung. Ecol. 39, 336–348 (2019).Article 

Google Scholar 
34.Boddy, L., Owens, E. M. & Chapela, I. H. Small scale variation in decay rate within logs one year after felling: effect of fungal community structure and moisture content. FEMS Microbiol. Ecol. 62, 173–184 (1989).Article 

Google Scholar 
35.Setälä, H. & McLean, M. A. Decomposition rate of organic substrates in relation to the species diversity of soil saprophytic fungi. Oecologia 139, 98–107 (2004).PubMed 
Article 
ADS 
PubMed Central 

Google Scholar 
36.Yang, C. et al. Higher fungal diversity is correlated with lower CO2 emissions from dead wood in a natural forest. Sci. Rep. 6, 31066 (2016).CAS 
PubMed 
PubMed Central 
Article 
ADS 

Google Scholar 
37.Berg, B., & McClaugherty, C. Plant litter: Decomposition, humus formation, carbon sequestration (Springer, 2003).38.Fukasawa, Y., Takahashi, K., Arikawa, T., Hattori, T. & Maekawa, N. Fungal wood decomposer activities influence community structure of myxomycetes and bryophytes on coarse woody debris. Fung. Ecol. 14, 44–52 (2015).Article 

Google Scholar 
39.Fukasawa, Y., Hyodo, F. & Kawakami, S. Foraging association between myxomycetes and fungal communities on coarse woody debris. Soil Biol. Biochem. 121, 95–102 (2018).CAS 
Article 

Google Scholar 
40.Fukasawa, Y. Fungal succession and decomposition of Pinus densiflora snags. Ecol. Res. 33, 435–444 (2018).Article 

Google Scholar 
41.Fukasawa, Y., Osono, T. & Takeda, H. Effects of attack of saprobic fungi on twig litter decomposition by endophytic fungi. Ecol. Res. 24, 1067–1073 (2009).Article 

Google Scholar 
42.Hiscox, J. & Boddy, L. Armed and dangerous—Chemical warfare in wood decay communities. Fung. Biol. Rev. 31, 169–184 (2017).Article 

Google Scholar 
43.Presley, G.N., Zhang, J., Purvine, S.O., & Schilling, J.S. Functional genomics, transcriptomics, and proteomics reveal distinct combat strategies between lineages of wood-degrading fungi with redundant wood decay mechanisms. Front. Microbiol. 11, article 1646 (2020).44.Hiscox, J., Savoury, M., Vaughan, I. P., Muller, C. T. & Boddy, L. Antagonistic fungal interactions influence carbon dioxide evolution from decomposing wood. Fung. Ecol. 14, 24–32 (2015).Article 

Google Scholar 
45.Zhang, X., Xu, C. & Wang, H. Pretreatment of bamboo residues with Coriolus versicolor for enzymatic hydrolysis. J. Biosci. Bioengineer. 104, 149–151 (2007).CAS 
Article 

Google Scholar 
46.Horisawa, S., Inoue, A. & Yamanaka, Y. Direct ethanol production from lignocellulosic materials by mixed culture of wood rot fungi Schizophyllum commune, Bjerkandera adusta, and Fomitopsis palustris. Fermentation 5, 21 (2019).CAS 
Article 

Google Scholar 
47.Schilling, J. S., Kaffenberger, J. T., Held, B. W., Ortiz, R. & Blanchette, R. A. Using wood rot phenotypes to illuminate the “Gray” among decomposer fungi. Front. Microbiol 11, 1288 (2020).PubMed 
PubMed Central 
Article 

Google Scholar 
48.Crawford, R. H., Carpenter, S. E. & Harmon, M. E. Communities of filamentous fungi and yeast in decomposing logs of Pseudotsuga menziesii. Mycologia 82, 759–765 (1990).Article 

Google Scholar 
49.Lumley, T. C., Gignac, L. D. & Currah, R. S. Microfungus communities of white spruce and trembling aspen logs and different stages of decay in disturbed and undisturbed sites in the boreal mixedwood region of Alberta. Can. J. Bot. 79, 76–92 (2001).
Google Scholar 
50.Fukasawa, Y., Osono, T. & Takeda, H. Microfungus communities of Japanese beech logs at different stages of decay in a cool temperate deciduous forest. Can. J. For. Res. 39, 1606–1614 (2009).CAS 
Article 

Google Scholar 
51.Fukasawa, Y., Osono, T. & Takeda, H. Dynamics of physicochemical properties and occurrence of fungal fruit bodies during decomposition of coarse woody debris of Fagus crenata. J. For. Res. 14, 20–29 (2009).CAS 
Article 

Google Scholar 
52.Maynard, D., Crowther, T. W. & Bradford, M. A. Competitive network determines the direction of the diversity-function relationship. Proc. Natl. Acad. Sci. USA 114, 11464–11469 (2017).CAS 
PubMed 
Article 

Google Scholar 
53.Kubart, A., Vasaitis, R., Stenlid, J. & Dahlberg, A. Fungal communities in Norway spruce stumps along a latitudinal gradient in Sweden. For. Ecol. Manag. 371, 50–58 (2016).Article 

Google Scholar 
54.MacArthur, R.H., & Wilson, E.O. The Theory of Island Biogeography. (Princeton University Press, 2001).55.Yachi, S. & Loreau, M. Biodiversity and ecosystem functioning productivity in a fluctuating environment: The insurance hypothesis. Proc. Natl. Acad. Sci. USA 96, 1463–1468 (1999).CAS 
PubMed 
Article 
ADS 

Google Scholar 
56.Maynard, D. et al. Consistent trade-offs in fungal trait expression across broad spatial scales. Nat. Microbiol. 4, 846–853 (2019).CAS 
PubMed 
Article 

Google Scholar 
57.Talbot, J. M. et al. Endemism and functional convergence across the North American soil mycobiome. Proc. Natl. Acad. Sci. USA 111, 6341–6346 (2014).CAS 
PubMed 
Article 
ADS 

Google Scholar 
58.Pyle, C. & Brown, M. M. Heterogeneity of wood decay classes within hardwood logs. For. Ecol. Manag. 114, 253–259 (1999).Article 

Google Scholar 
59.Carini, P. et al. Effects of spatial variability and relic DNA removal on the detection of temporal dynamics in soil microbial communities. mBio 11, e02776-19 (2020).60.Fukasawa, Y. & Matsuoka, S. Communities of wood-inhabiting fungi in dead pine logs along a geographical gradient in Japan. Fung. Ecol. 18, 75–82 (2015).Article 

Google Scholar 
61.Worrall, J. J., Anagnost, S. E. & Zabel, R. A. Comparison of wood decay among diverse lignicolous fungi. Mycologia 89, 199–219 (1997).Article 

Google Scholar 
62.Deacon, J. W. Decomposition of filter paper cellulose by thermophilic fungi acting singly, in combination, and in sequence. Tr. Br. Mycol. Soc. 85, 663–669 (1985).CAS 
Article 

Google Scholar 
63.Fukasawa, Y. Effects of wood decomposer fungi on tree seedling establishment on coarse woody debris. For. Ecol. Manag. 266, 232–238 (2012).Article 

Google Scholar 
64.Toju, H., Tanabe, A. S., Yamamoto, S. & Sato, H. High-coverage ITS primers for the DNA-based identification of ascomycetes and basidiomycetes in environmental samples. PLoS ONE 7, e40863 (2012).CAS 
PubMed 
PubMed Central 
Article 
ADS 

Google Scholar 
65.Tanabe, A. S. & Toju, H. Two new computational methods for universal DNA barcoding: A benchmark using barcode sequences of bacteria, archaea, animals, fungi, and land plants. PLoS ONE 8, e76910 (2013).CAS 
PubMed 
PubMed Central 
Article 
ADS 

Google Scholar 
66.Osono, T. Metagenomic approach yields insights into fungal diversity and functioning. In Species diversity and community structure (eds Sota, T., Kagata, H., Ando, Y., Utsumi, S., & Osono, T.) 1–23 (Springer, 2014).67.Ohtsubo, Y., Ikeda-Ohtsubo, W., Nagata, Y. & Tsuda, M. GenomeMatcher: a graphical user interface for DNA sequence comparison. BMC Bioinform. 9, 376. https://doi.org/10.1186/1471-2105-9-376 (2008).CAS 
Article 

Google Scholar 
68.Ovaskainen, O., & Abrego, N. Joint Species Distribution Modelling: With Application in R (Cambridge University Press, 2020).69.R Core Team. R: A language and environment for statistical computing. The R Foundation for Statistical Computing, Vienna, Austria. www.R-project.org (2019). More

1.Stal, L. J. et al. BASIC: Baltic Sea cyanobacteria. An investigation of the structure and dynamics of water blooms of cyanobacteria in the Baltic Sea—Responses to a changing environment. Cont. Shelf Res. 23, 1695–1714 (2003).Article 
ADS 

Google Scholar 
2.McGregor, G. B. et al. First report of a toxic Nodularia spumigena (nostocales/cyanobacteria) bloom in sub-tropical Australia. I. Phycological and public health investigations. Int. J. Env. Res. Public Health 9, 2396–2411 (2012).Article 

Google Scholar 
3.Popin, R. V. et al. Genomic and metabolomic analyses of natural products in Nodularia spumigena isolated from a shrimp culture pond. Toxins 12, 141 (2020).CAS 
PubMed Central 
Article 
PubMed 

Google Scholar 
4.Seaman, M., Ashton, P. & Williams, W. Inland salt waters of southern Africa. Hydrobiologia 210, 75–91 (1991).CAS 
Article 

Google Scholar 
5.Beutel, M. W., Horne, A. J., Roth, J. C. & Barratt, N. J. Saline Lakes 91–105 (Springer, 2001).Book 

Google Scholar 
6.Paerl, H. W. & Paul, V. J. Climate change: Links to global expansion of harmful cyanobacteria. Water Res. 46, 1349–1363 (2012).CAS 
PubMed 
Article 

Google Scholar 
7.Karjalainen, M. et al. Ecosystem consequences of cyanobacteria in the northern Baltic Sea. AMBIO J. Human Environ. 36, 195–202 (2007).CAS 
Article 

Google Scholar 
8.Sotton, B., Domaizon, I., Anneville, O., Cattanéo, F. & Guillard, J. Nodularin and cylindrospermopsin: A review of their effects on fish. Rev. Fish Biol. Fish. 25, 1–19 (2015).Article 

Google Scholar 
9.Mazur-Marzec, H., Bertos-Fortis, M., Toruńska-Sitarz, A., Fidor, A. & Legrand, C. Chemical and genetic diversity of Nodularia spumigena from the Baltic Sea. Mar. Drugs 14, 209. https://doi.org/10.3390/md14110209 (2016).CAS 
Article 
PubMed Central 
PubMed 

Google Scholar 
10.Voss, B. et al. Insights into the physiology and ecology of the brackish-water-adapted Cyanobacterium Nodularia spumigena CCY9414 based on a genome-transcriptome analysis. PLoS ONE 8, e60224–e60224. https://doi.org/10.1371/journal.pone.0060224 (2013).CAS 
Article 
PubMed 
PubMed Central 
ADS 

Google Scholar 
11.Le Manach, S. et al. Global metabolomic characterizations of Microcystis spp. highlights clonal diversity in natural bloom-forming populations and expands metabolite structural diversity. Front. Microbiol. https://doi.org/10.3389/fmicb.2019.00791 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
12.Welker, M. & von Döhren, H. Cyanobacterial peptides—Nature’s own combinatorial biosynthesis. FEMS Microbiol. Rev. 30, 530–563 (2006).CAS 
PubMed 
Article 

Google Scholar 
13.Kehr, J. C., Picchi, D. G. & Dittmann, E. Natural product biosyntheses in cyanobacteria: A treasure trove of unique enzymes. Beilstein J. Org. Chem. 7, 1622–1635. https://doi.org/10.3762/bjoc.7.191 (2011).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
14.Christiansen, G., Philmus, B., Hemscheidt, T. & Kurmayer, R. Genetic variation of adenylation domains of the anabaenopeptin synthesis operon and evolution of substrate promiscuity. J. Bacteriol. 193, 3822–3831 (2011).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
15.Ishida, K. et al. Biosynthesis and structure of aeruginoside 126A and 126B, cyanobacterial peptide glycosides bearing a 2-carboxy-6-hydroxyoctahydroindole moiety. Chem. Biol. 14, 565–576 (2007).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
16.Fewer, D.P. et al. The non-ribosomal assembly and frequent occurrence of the protease inhibitors spumigins in the bloom-forming cyanobacterium Nodularia spumigena. Mol. Microbiol. 73, 924–937. https://doi.org/10.1111/j.1365-2958.2009.06816.x (2009).CAS 
Article 
PubMed 

Google Scholar 
17.Portmann, C. et al. Isolation of aerucyclamides C and D and structure revision of microcyclamide 7806A: Heterocyclic ribosomal peptides from Microcystis aeruginosa PCC 7806 and their antiparasite evaluation. J. Nat. Prod. 71, 1891–1896 (2008).CAS 
PubMed 
Article 

Google Scholar 
18.Ersmark, K., Del Valle, J. R. & Hanessian, S. Chemistry and biology of the aeruginosin family of serine protease inhibitors. Angew. Chem. Int. Ed. 47, 1202–1223 (2008).CAS 
Article 

Google Scholar 
19.Liu, L. et al. Pseudoaeruginosins, nonribosomal peptides in Nodularia spumigena. ACS Chem. Biol. 10, 725–733 (2015).CAS 
PubMed 
Article 

Google Scholar 
20.Itou, Y., Suzuki, S., Ishida, K. & Murakami, M. Anabaenopeptins G and H, potent carboxypeptidase A inhibitors from the cyanobacterium Oscillatoria agardhii (NIES-595). Bioorg. Med. Chem. Lett. 9, 1243–1246 (1999).CAS 
PubMed 
Article 

Google Scholar 
21.Bister, B. et al. Cyanopeptolin 963A, a chymotrypsin inhibitor of Microcystis PCC 7806. J. Nat. Prod. 67, 1755–1757 (2004).CAS 
PubMed 
Article 

Google Scholar 
22.Neilan, B. A., Pearson, L. A., Muenchhoff, J., Moffitt, M. C. & Dittmann, E. Environmental conditions that influence toxin biosynthesis in cyanobacteria. Environ. Microbiol. 15, 1239–1253. https://doi.org/10.1111/j.1462-2920.2012.02729.x (2013).CAS 
Article 
PubMed 

Google Scholar 
23.Halstvedt, C. B., Rohrlack, T., Ptacnik, R. & Edvardsen, B. On the effect of abiotic environmental factors on production of bioactive oligopeptides in field populations of Planktothrix spp. (Cyanobacteria). J. Plankton Res. 30, 607–617 (2008).CAS 
Article 

Google Scholar 
24.Mazur-Marzec, H. et al. Diversity of peptides produced by Nodularia spumigena from various geographical regions. Mar. Drugs 11, 1–19. https://doi.org/10.3390/md11010001 (2012).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
25.Repka, S., Koivula, M., Harjunpa, V., Rouhiainen, L. & Sivonen, K. Effects of phosphate and light on growth of and bioactive peptide production by the Cyanobacterium anabaena strain 90 and its anabaenopeptilide mutant. Appl. Environ. Microbiol. 70, 4551–4560. https://doi.org/10.1128/aem.70.8.4551-4560.2004 (2004).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
26.Lehtimäki, J., Moisander, P., Sivonen, K. & Kononen, K. Growth, nitrogen fixation, and nodularin production by two Baltic Sea cyanobacteria. Appl. Environ. Microbiol. 63, 1647–1656 (1997).PubMed 
PubMed Central 
Article 

Google Scholar 
27.OECD. Test No. 201: Freshwater alga and cyanobacteria, growth inhibition test. OECD Guidelines for the Testing of Chemicals, Section 2. https://doi.org/10.1787/9789264069923-en (OECD
Publishing, Paris, 2011).28.Vaas, L. A. I., Sikorski, J., Michael, V., Göker, M. & Klenk, H.-P. Visualization and curve-parameter estimation strategies for efficient exploration of phenotype microarray kinetics. PLoS ONE 7, e34846. https://doi.org/10.1371/journal.pone.0034846 (2012).CAS 
Article 
PubMed 
PubMed Central 
ADS 

Google Scholar 
29.Higo, S., Yamatogi, T., Ishida, N., Hirae, S. & Koike, K. Application of a pulse-amplitude-modulation (PAM) fluorometer reveals its usefulness and robustness in the prediction of Karenia mikimotoi blooms: A case study in Sasebo Bay, Nagasaki, Japan. Harmful Algae 61, 63–70 (2017).Article 

Google Scholar 
30.Qi, H., Wang, J. & Wang, Z. A comparative study of maximal quantum yield of photosystem II to determine nitrogen and phosphorus limitation on two marine algae. J. Sea Res. 80, 1–11 (2013).Article 
ADS 

Google Scholar 
31.Briand, E., Bormans, M., Gugger, M., Dorrestein, P. C. & Gerwick, W. H. Changes in secondary metabolic profiles of Microcystis aeruginosa strains in response to intraspecific interactions. Environ. Microbiol. 18, 384–400. https://doi.org/10.1111/1462-2920.12904 (2016).CAS 
Article 
PubMed 

Google Scholar 
32.Koek, M. M., Muilwijk, B., van der Werf, M. J. & Hankemeier, T. Microbial metabolomics with gas chromatography/mass spectrometry. Anal. Chem. 78, 1272–1281 (2006).CAS 
PubMed 
Article 

Google Scholar 
33.R Development Core Team. R: A language and environment for statistical computing. (R Foundation for Statistical Computing, Vienna, Austria, 2020).34.Medlock, G. L. et al. Inferring metabolic mechanisms of interaction within a defined gut microbiota. Cell Syst. 7, 245-257.e247. https://doi.org/10.1016/j.cels.2018.08.003 (2018).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
35.Paul, C., Mausz, M. A. & Pohnert, G. A co-culturing/metabolomics approach to investigate chemically mediated interactions of planktonic organisms reveals influence of bacteria on diatom metabolism. Metabolomics 9, 349–359. https://doi.org/10.1007/s11306-012-0453-1 (2013).CAS 
Article 

Google Scholar 
36.Schatz, D. et al. Ecological implications of the emergence of non-toxic subcultures from toxic Microcystis strains. Environ. Microbiol. 7, 798–805 (2005).CAS 
PubMed 
Article 

Google Scholar 
37.Jensen, A., Rystad, B. & Skoglund, L. The use of dialysis culture in phytoplankton studies. J. Exp. Mar. Biol. Ecol. 8, 241–248 (1972).Article 

Google Scholar 
38.Kobayashi, K., Takata, Y. & Kodama, M. Direct contact between Pseudo-nitzschiaámultiseries and bacteria is necessary for the diatom to produce a high level of domoic acid. Fish. Sci. 75, 771–776 (2009).CAS 
Article 

Google Scholar 
39.McVeigh, I., & Brown, W. In vitro growth of chlamydomonas chlamydogama bold and haematococcus pluvialis flotow em. Wille in mixed cultures.
Bulletin of the Torrey Botanical Club, 81(3), 218–233. https://doi.org/10.2307/2481813 (1954).CAS 
Article 

Google Scholar 
40.Sieg, R. D., Poulson-Ellestad, K. L. & Kubanek, J. Chemical ecology of the marine plankton. Nat. Prod. Rep. 28, 388–399 (2011).CAS 
PubMed 
Article 

Google Scholar 
41.Yamasaki, A. An overview of CO2 mitigation options for global warming—Emphasizing CO2 sequestration options. J. Chem. Eng. Japan 36, 361–375 (2003).CAS 
Article 

Google Scholar 
42.Hajdu, S., Hoglander, H. & Larsson, U. Phytoplankton vertical distributions and composition in Baltic Sea cyanobacterial blooms. Harmful Algae 6, 189–205 (2007).Article 

Google Scholar 
43.Berman-Frank, I. & Dubinsky, Z. Balanced growth in aquatic plants: Myth or reality? Phytoplankton use the imbalance between carbon assimilation and biomass production to their strategic advantage. Bioscience 49, 29–37 (1999).Article 

Google Scholar 
44.Kruskopf, M. & Flynn, K. J. Chlorophyll content and fluorescence responses cannot be used to gauge reliably phytoplankton biomass, nutrient status or growth rate. New Phytol. 169, 525–536. https://doi.org/10.1111/j.1469-8137.2005.01601.x (2006).CAS 
Article 
PubMed 

Google Scholar 
45.Raven, J. A. & Beardall, J. Chlorophyll fluorescence and ecophysiology: Seeing red?. New Phytol. 169, 449–451. https://doi.org/10.1111/j.1469-8137.2006.01637.x (2006).CAS 
Article 
PubMed 

Google Scholar 
46.Li, Q. et al. A large-scale comparative metagenomic study reveals the functional interactions in six bloom-forming microcystis-epibiont communities. Front. Microbiol. https://doi.org/10.3389/fmicb.2018.00746 (2018).Article 
PubMed 
PubMed Central 

Google Scholar 
47.Harke, M. J. et al. A review of the global ecology, genomics, and biogeography of the toxic cyanobacterium Microcystis spp. Harmful Algae 54, 4–20 (2016).PubMed 
Article 

Google Scholar 
48.Caldwell, D. Associations between photosynthetic and heterotrophic prokaryotes in plankton. in Abstracts of the third International Symposium on Photosynthetic Prokaryotes (ed Nichols, J. M) (University of Liverpool, UK, 1979).49.Park, H. D. et al. Degradation of the cyanobacterial hepatotoxin microcystin by a new bacterium isolated from a hypertrophic lake. Environ. Toxicol. Int. J. 16, 337–343 (2001).CAS 
Article 
ADS 

Google Scholar 
50.Berg, C. et al. Dissection of microbial community functions during a cyanobacterial bloom in the Baltic Sea via metatranscriptomics. Front. Mar. Sci. 5, 55 (2018).Article 

Google Scholar 
51.Humbert, J.-F. et al. A tribute to disorder in the genome of the bloom-forming freshwater cyanobacterium Microcystis aeruginosa. PLoS ONE 8, e70747. https://doi.org/10.1371/journal.pone.0070747 (2013).CAS 
Article 
PubMed 
PubMed Central 
ADS 

Google Scholar 
52.Toporowska, M., Mazur-Marzec, H. & Pawlik-Skowrońska, B. The effects of cyanobacterial bloom extracts on the biomass, Chl-a, MC and other oligopeptides contents in a natural Planktothrix agardhii population. Int. J. Env. Res. Public Health 17, 2881 (2020).CAS 
Article 

Google Scholar 
53.Grabowska, M., Kobos, J., Toruńska-Sitarz, A. & Mazur-Marzec, H. Non-ribosomal peptides produced by Planktothrix agardhii from Siemianówka Dam Reservoir SDR (northeast Poland). Arch. Microbiol. 196, 697–707 (2014).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
54.Penn, K., Wang, J., Fernando, S. C. & Thompson, J. R. Secondary metabolite gene expression and interplay of bacterial functions in a tropical freshwater cyanobacterial bloom. ISME J. 8, 1866–1878 (2014).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
55.Neilan, B. A. et al. Nonribosomal peptide synthesis and toxigenicity of cyanobacteria. J. Bacteriol. 181, 4089–4097 (1999).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
56.Long, B. M., Jones, G. J. & Orr, P. T. Cellular microcystin content in N-limited Microcystis aeruginosa can be predicted from growth rate. Appl. Environ. Microbiol. 67, 278–283 (2001).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
57.Qu, J. et al. Determination of the role of microcystis aeruginosa in toxin generation based on phosphoproteomic profiles. Toxins 10, 304 (2018).PubMed Central 
Article 
CAS 
PubMed 

Google Scholar 
58.Raven, J. A. Cyanotoxins: A poison that frees phosphate. Curr. Biol. 20, R850–R852 (2010).CAS 
PubMed 
Article 

Google Scholar 
59.Utkilen, H. & Gjølme, N. Iron-stimulated toxin production in Microcystis aeruginosa. Appl. Environ. Microbiol. 61, 797–800 (1995).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
60.Gan, N. et al. The role of microcystins in maintaining colonies of bloom-forming Microcystis spp. Environ. Microbiol. 14, 730–742 (2012).CAS 
PubMed 
Article 

Google Scholar 
61.Pomati, F., Rossetti, C., Manarolla, G., Burns, B. P. & Neilan, B. A. Interactions between intracellular Na+ levels and saxitoxin production in Cylindrospermopsis raciborskii T3. Microbiology 150, 455–461 (2004).CAS 
PubMed 
Article 

Google Scholar 
62.Seigler, D. & Price, P. W. Secondary compounds in plants: Primary functions. Am. Nat. 110, 101–105 (1976).CAS 
Article 

Google Scholar 
63.Zilliges, Y. et al. The cyanobacterial hepatotoxin microcystin binds to proteins and increases the fitness of Microcystis under oxidative stress conditions. PLoS ONE 6, e17615 (2011).CAS 
PubMed 
PubMed Central 
Article 
ADS 

Google Scholar 
64.Meissner, S., Fastner, J. & Dittmann, E. Microcystin production revisited: Conjugate formation makes a major contribution. Environ. Microbiol. 15, 1810–1820. https://doi.org/10.1111/1462-2920.12072 (2013).CAS 
Article 
PubMed 

Google Scholar 
65.Orr, P. T., Willis, A. & Burford, M. A. Application of first order rate kinetics to explain changes in bloom toxicity—The importance of understanding cell toxin quotas. J. Oceanol. Limnol. 36, 1063–1074. https://doi.org/10.1007/s00343-019-7188-z (2018).CAS 
Article 
ADS 

Google Scholar 
66.Rantala, A. et al. Phylogenetic evidence for the early evolution of microcystin synthesis. Proc. Natl. Acad. Sci. USA 101, 568–573 (2004).CAS 
PubMed 
Article 
ADS 

Google Scholar 
67.Orr, P. T. & Jones, G. J. Relationship between microcystin production and cell division rates in nitrogen-limited Microcystis aeruginosa cultures. Limnol. Oceanogr. 43, 1604–1614 (1998).CAS 
Article 
ADS 

Google Scholar 
68.Burford, M. A. et al. Understanding the winning strategies used by the bloom-forming cyanobacterium Cylindrospermopsis raciborskii. Harmful Algae 54, 44–53 (2016).PubMed 
Article 

Google Scholar 
69.Pierangelini, M. et al. Constitutive cylindrospermopsin pool size in Cylindrospermopsis raciborskii under different light and CO2 partial pressure conditions. Appl. Environ. Microbiol. 81, 3069–3076. https://doi.org/10.1128/aem.03556-14 (2015).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
70.Falkowski, P. G., Sukenik, A. & Herzig, R. Nitrogen limitation in Isochrysis galbana (Haptophyceae). II. Relative abundance of chloroplast proteins. J. Phycol. 25, 471–478 (1989).CAS 
Article 

Google Scholar 
71.Turpin, D. H. Effects of inorganic N availability on algal photosynthesis and carbon metabolism. J. Phycol. 27, 14–20 (1991).CAS 
Article 

Google Scholar 
72.Moffitt, M. C. & Neilan, B. A. Characterization of the nodularin synthetase gene cluster and proposed theory of the evolution of cyanobacterial hepatotoxins. Appl. Environ. Microbiol. 70, 6353–6362 (2004).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
73.Fewer, D. P. et al. New structural variants of aeruginosin produced by the toxic bloom forming cyanobacterium Nodularia spumigena. PLoS ONE 8, e73618 (2013).CAS 
PubMed 
PubMed Central 
Article 
ADS 

Google Scholar 
74.Fujii, K. et al. Comparative study of toxic and non-toxic cyanobacterial products: Novel peptides from toxic Nodularia spumigena AV1. Tetrahedron Lett. 38, 5525–5528 (1997).CAS 
Article 

Google Scholar 
75.Ishida, K. et al. Plasticity and evolution of aeruginosin biosynthesis in cyanobacteria. Appl. Environ. Microbiol. 75, 2017–2026 (2009).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
76.Suikkanen, S., Fistarol, G. O. & Granéli, E. Allelopathic effects of the Baltic cyanobacteria Nodularia spumdigena, Aphanizomenon flosaquae and Anabaena lemmermannii on algal monocultures. J. Exp. Mar. Biol. Ecol. 308, 85–101 (2004).Article 

Google Scholar 
77.Suikkanen, S., Engström-Öst, J., Jokela, J., Sivonen, K. & Viitasalo, M. Allelopathy of Baltic Sea cyanobacteria: No evidence for the role of nodularin. J. Plankton Res. 28, 543–550. https://doi.org/10.1093/plankt/fbi139 (2006).CAS 
Article 

Google Scholar 
78.Żak, A. & Kosakowska, A. The influence of extracellular compounds produced by selected Baltic cyanobacteria, diatoms and dinoflagellates on growth of green algae Chlorella vulgaris. Estuar. Coast. Shelf Sci. 167, 113–118 (2015).Article 
ADS 

Google Scholar 
79.Śliwińska-Wilczewska, S., Felpeto, A. B., Możdżeń, K., Vasconcelos, V. & Latała, A. Physiological effects on coexisting microalgae of the allelochemicals produced by the bloom-forming cyanobacteria Synechococcus sp. and Nodularia spumigena. Toxins 11, 712 (2019).PubMed Central 
Article 
CAS 
PubMed 

Google Scholar 
80.Gross, E. M. Allelopathy of aquatic autotrophs. Crit. Rev. Plant Sci. 22, 313–339 (2003).Article 

Google Scholar 
81.Legrand, C., Rengefors, K., Fistarol, G. O. & Graneli, E. Allelopathy in phytoplankton-biochemical, ecological and evolutionary aspects. Phycologia 42, 406–419 (2003).Article 

Google Scholar 
82.Leao, P. N., Vasconcelos, M. T. & Vasconcelos, V. M. Allelopathy in freshwater cyanobacteria. Crit. Rev. Microbiol. 35, 271–282. https://doi.org/10.3109/10408410902823705 (2009).CAS 
Article 
PubMed 

Google Scholar 
83.MacKintosh, C., Beattie, K. A., Klumpp, S., Cohen, P. & Codd, G. A. Cyanobacterial microcystin-LR is a potent and specific inhibitor of protein phosphatases 1 and 2A from both mammals and higher plants. FEBS Lett. 264, 187–192 (1990).CAS 
PubMed 
Article 

Google Scholar 
84.Pflugmacher, S. Possible allelopathic effects of cyanotoxins, with reference to microcystin-LR, in aquatic ecosystems. Environ. Toxicol. Int. J. 17, 407–413 (2002).CAS 
Article 
ADS 

Google Scholar 
85.Tilahun S. Exclusive partitioning of intra- and extra-cellular cyanotoxins: limitation of the conventional procedure. Environ. Sci. Pollut. Res. Int. 27(14), 17427–17428. https://doi.org/10.1007/s11356-020-08256-8 (2020).Article 
PubMed 

Google Scholar 
86.Park, H. D. et al. Temporal variabilities of the concentrations of intra-and extracellular microcystin and toxic Microcystis species in a hypertrophic lake, Lake Suwa, Japan (1991–1994). Environ. Toxicol. Water Qual. Int. J. 13, 61–72 (1998).CAS 
Article 
ADS 

Google Scholar 
87.Tsuji, K. et al. Stability of microcystins from cyanobacteria: Effect of light on decomposition and isomerization. Environ. Sci. Technol. 28, 173–177 (1994).CAS 
PubMed 
Article 
ADS 
PubMed Central 

Google Scholar 
88.Schatz, D. et al. Towards clarification of the biological role of microcystins, a family of cyanobacterial toxins. Environ. Microbiol. 9, 965–970 (2007).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
89.Makower, A. K. et al. Transcriptomics-aided dissection of the intracellular and extracellular roles of microcystin in Microcystis aeruginosa PCC 7806. Appl. Environ. Microbiol. 81, 544–554 (2015).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
90.Kaplan, A. et al. The languages spoken in the water body (or the biological role of cyanobacterial toxins). Front. Microbiol. 3, 138 (2012).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
91.Svercel, M. Negative allelopathy among cyanobacteria. in Cyanobacteria: Ecology, Toxicology and Management. (ed Ferrao-Filho, A. S.) 27–46 (Nova Science Publishers, New York, NY, USA, 2013).
Google Scholar 
92.Wiegand, C. & Pflugmacher, S. Ecotoxicological effects of selected cyanobacterial secondary metabolites a short review. Toxicol. Appl. Pharmacol. 203, 201–218. https://doi.org/10.1016/j.taap.2004.11.002 (2005).CAS 
Article 
PubMed 

Google Scholar 
93.Agrawal, M. & Agrawal, M. K. Cyanobacteria–herbivore interaction in freshwater ecosystem. J. Microbiol. Biotechnol. Res. 1, 52–66 (2011).
Google Scholar 
94.Sadler, T. & von Elert, E. Dietary exposure of Daphnia to microcystins: No in vivo relevance of biotransformation. Aquat. Toxicol. 150, 73–82. https://doi.org/10.1016/j.aquatox.2014.02.017 (2014).CAS 
Article 
PubMed 

Google Scholar 
95.Rohrlack, T., Christiansen, G. & Kurmayer, R. Putative antiparasite defensive system involving ribosomal and nonribosomal oligopeptides in cyanobacteria of the genus Planktothrix. Appl. Environ. Microbiol. 79, 2642–2647 (2013).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
96.Sivonen, K. et al. Occurrence of the hepatotoxic cyanobacterium Nodularia spumigena in the Baltic Sea and structure of the toxin. Appl. Environ. Microbiol. 55, 1990–1995 (1989).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
97.Burbage, C. D. & Binder, B. J. Relationship between cell cycle and light-limited growth rate in oceanic Prochlorococcus (MIT9312) and Synechococcus (WH8103) (Cyanobacteria). J. Phycol. 43, 266–274. https://doi.org/10.1111/j.1529-8817.2007.00315.x (2007).Article 

Google Scholar 
98.Lei, L., Dai, J., Lin, Q., Peng, L. Competitive dominance of Microcystis aeruginosa against Raphidiopsis raciborskii is strain-and temperature dependent. Knowl. Manag. Aquat. Ecosyst. 421, 36. https://doi.org/10.1051/kmae/2020023 (2020).Article 

Google Scholar  More

1.Schulthess, C. P. & Huang, C. P. Adsorption of heavy metals by silicon and aluminum oxide surfaces on clay minerals. Soil Sci. Soc. Am. J. 54, 679–688 (1990).ADS 
Article 

Google Scholar 
2.City of Joplin Health Department. Report to Jasper County EPA Superfund citizen’s task force. City of Joplin Health Department, Joplin, MO (1995).3.Beyer, W. N., Pattee, O. H., Sileo, L., Hoffman, D. J. & Mulhern, B. M. Metal contamination in wildlife living near two zinc smelters. Environ. Pollut. Ser. A38, 63–86 (1985).Article 

Google Scholar 
4.Khan, D. H. & Frankland, B. Effects of cadmium and lead on radish plants with particular reference to movement of metals through soil profile and plant. Plant Soil 70, 335–345 (1983).CAS 
Article 

Google Scholar 
5.Ruby, M. V., Davis, A., Kempton, J. H., Drexler, J. & Bergstrom, P. D. Lead bioavailability: Dissolution kinetics under simulated gastric conditions. Environ. Sci. Technol. 26, 1242–1248 (1992).ADS 
CAS 
Article 

Google Scholar 
6.Ruby, M. V., Davis, A. & Nicholson, A. In-situ formation of lead phosphates in soils as a method to immobilize lead. Environ. Sci. Tchnol. 28, 646–654 (1994).ADS 
CAS 
Article 

Google Scholar 
7.Nriagu, J. O. Lead orthophosphates-IV: Formation and stability in the environment. Geochim. Cosmochim. Acta 38, 887–898 (1974).ADS 
CAS 
Article 

Google Scholar 
8.Ma, Q. Y., Logan, T. J. & Traina, S. J. Lead immobilization from aqueous solutions and contaminated soils using phosphate rocks. Environ. Sci. Technol. 27, 1118–1126 (1995).ADS 
Article 

Google Scholar 
9.Xu, Y. & Schwartz, F. W. Lead immobilization by hydroxyapatite in aqueous solutions. J. Contam. Hydrol. 15, 187–206 (1994).ADS 
CAS 
Article 

Google Scholar 
10.Zhang, P. C., Ryan, J. A. & Yang, J. In vitro soil Pb solubility in the presence of hydroxyapatite. Environ. Sci. Technol. 32, 2763–2768 (1998).ADS 
CAS 
Article 

Google Scholar 
11.Osborne, L. R., Baker, L. L. & Strawn, D. G. Lead immobilization and phosphorus availability in phosphate-amended, mine-contaminated soils. J. Environ. Qual. 44(1), 183–190 (2015).Article 

Google Scholar 
12.Yang, J. J., Mosby, D. E., Casteel, S. W. & Blanchar, R. W. Lead immobilization using phosphoric acid in smelter-contaminated urban soil. Environ. Sci. Technol. 35, 3553–3559 (2001).ADS 
CAS 
Article 

Google Scholar 
13.Tang, X., Yang, J., Goyne, K. W. & Deng, B. Long-term risk reduction of lead contaminated urban soil by phosphate treatment. Env. Eng. Sci. 26(12), 1747–1754 (2009).CAS 
Article 

Google Scholar 
14.Tang, X. & Yang, J. Long-term stability of risk assessment of lead mill waste treated by soluble phosphate. Sci. Total Environ. 438, 299–303 (2012).ADS 
CAS 
Article 

Google Scholar 
15.Brown, S. L. & Chaney, R. L. A rapid in-vitro procedure to a mimicked in-situ remediation study of metal-contaminated soils characterize the effectiveness of a variety of in-situ lead remediation with emphasis on Cd and Pb. J. Environ. Qual. 23, 58–63 (1997).
Google Scholar 
16.Li, Y. M., Chaney, R. L., Siebielec, G. & Kerschner, B. Response of four turf grass cultivars to limestone and biosolids-compost amendment of a zinc and cadmium contaminated soil at Palmerton, Pennsylvania. J. Environ. Qual. 29, 1440–1447 (2000).CAS 
Article 

Google Scholar 
17.Basta, N. T., Gradwhol, R., Snethen, K. L. & Schroder, J. L. Chemical immobilization of lead, zinc and cadmium in smelter-contaminated soils using biosolids and rock phosphate. J. Environ. Qual. 30, 1222–1230 (2001).CAS 
Article 

Google Scholar 
18.Brown, S., Chaney, R. L., Hallfrisch, J. G. & Xue, Q. Effects of biosolids processing on lead bioavailability in an urban soil. J. Environ. Qual. 32, 100–108 (2003).CAS 
Article 

Google Scholar 
19.Mosby, D.E., Miller, S., Bishop, C., Mehuys, J. Former and abandoned lead and zinc mines demonstration project. Final Report. Missouri Department of Nature Resources, Jefferson City, MO (2002).20.Singh, S. P., Ma, L. Q., Tack, F. G. & Verloo, M. G. Trace metal leachability of land-disposed dredged sediments. J. Environ. Qual. 29, 1124–1142 (2000).CAS 
Article 

Google Scholar 
21.Yang, J., Tang, X. & Wang, Z. Y. Water quality and ecotoxicity as influenced by phosphate and biosolid treatments in lead-contaminated soil and mine waste. J. Environ. Monit. Rest. 3, 21–33 (2007).
Google Scholar 
22.Ma, L. Q. & Rao, G. N. Chemical fractionation of cadmium, copper, nickel, and zinc in contaminated soils. J. Environ. Qual. 26, 259–264 (1997).CAS 
Article 

Google Scholar 
23.Bhattacharyya, P., Chakrabarti, K., Chakraborty, A., Tripathy, S. & Powell, M. A. Fractionation and bioavailability of Pb in municipal solid waste compost and Pb uptake by rice straw and grain under submerged condition in amended soil. Geosci. J. 12, 41–45 (2008).ADS 
CAS 
Article 

Google Scholar 
24.Brady, N. & Weil, R. The Nature and Property of Soils (Prentice Hall, 2002).
Google Scholar 
25.Fu, H. et al. Cadmium and lead speciation as affected by soil amendments in calcareous soil. Environ. Eng. Sci. https://doi.org/10.1089/ees.2017.0307 (2017).Article 

Google Scholar 
26.Xu, J. C., Huang, L. M., Chen, C., Wang, J. & Long, X. X. Effective lead immobilization by phosphate rock solubilization mediated by phosphate rock amendment and phosphate solubilizing bacteria. Chemosphere 237, 124540 (2019).ADS 
CAS 
Article 

Google Scholar 
27.Kungolos, A. et al. Toxic properties of metals and organotin compounds and their interactions on daphnia magna and vibrio fischeri. Water Ai, Soil Pollut. 4, 101–110 (2004).CAS 
Article 

Google Scholar 
28.Eighmy, T. T. et al. Heavy metal stabilization in municipal solid waste combustion dry scrubber residue using soluble phosphate. Environ. Sci. Technol. 31, 3330–3338 (1997).ADS 
CAS 
Article 

Google Scholar 
29.Crannell, B. S. et al. Heavy metal stabilization in municipal solid waste combustion bottom ash using soluble phosphate. Waste Manag. 20, 135–148 (2000).CAS 
Article 

Google Scholar 
30.Bubb, J. M. & Lester, J. N. Impact of heavy metals on low land rivers and implications for the man and the environment. Sci. Total Environ. 100, 207–258 (1991).ADS 
CAS 
Article 

Google Scholar 
31.Zhang, M. K. et al. Solubility of phosphorus and heavy metals in potting media amended with yard waste-biosolids compost. J. Environ. Qual. 33, 373–379 (2004).CAS 
Article 

Google Scholar 
32.Xian, X. Effect of chemical forms of cadmium, zinc, and lead in polluted soils on their uptake by cabbage plants. Plant Soil 113, 257–264 (1989).CAS 
Article 

Google Scholar 
33.Zhang, M. K. et al. Phosphorus and heavy metal attachment and release in sandy soil aggregate fractions. Soil Sci. Soc. Am. J. 67, 1158–1167 (2003).ADS 
CAS 
Article 

Google Scholar 
34.Pierzynski, G. M. & Schwab, A. P. Bioavailability of zinc, cadmium, and lead in a metal contaminated alluvial soil. J. Environ. Qual. 22, 247–254 (1993).CAS 
Article 

Google Scholar 
35.Samaras, V., Tsadilas, C.D. Distribution and availability of six heavy metals in a soil treated with sewage sludge. In Proceedings of International Conference Biogeochemistry of Trace Elements, Berkeley, CA (1997).36.Scheckel, K. G. & Ryan, J. A. Spectroscopic speciation and quantification of lead in phosphate-amended soils. J. Environ. Qual. 33, 1288–1295 (2004).CAS 
Article 

Google Scholar 
37.Lang, F. & Kaupenjohann, M. Effect of dissolved organic matter on the precipitation and mobility of the lead compound chloropyromorphite in solution. Eur. J. Soil Sci. 54, 139–147 (2003).CAS 
Article 

Google Scholar 
38.Shi, Q. et al. Lead immobilization by phosphate in the presence of iron oxides: Adsorption versus precipitation. Water Res. 179, 115853 (2020).CAS 
Article 

Google Scholar 
39.Andrunik, M., Wolowiec, M., Wojnarski, D., Zelek-Pogudz, S. & Bajda, T. Transformation of Pb, Cd, and Zn minerals using phosphates. Minerals. 10, 342 (2020).CAS 
Article 

Google Scholar 
40.Guo, J. H. et al. Stablizing lead bullets in shooting range soil by phosphate-based surface coating. AIMS Environ Sci. 3(3), 474–487 (2016).CAS 
Article 

Google Scholar  More

This section summarizes how we track a population’s biological resource demand and domestic availability. We also explain which income metrics we chose. A more complete discussion of the resource metric method is included in the Supplementary Methods.Measuring the biological resource balanceThe sustainable development literature has consistently recognized the importance of biological resource security. For example, the foundational Brundtland report expressed it as the need to live “within the planet’s ecological means” or “in harmony with the changing productive potential of the ecosystem”43.These principles call for comparing biological resource regeneration with a population’s demand on nature. Since people’s demands compete for nature’s products and services, one way of measuring this relationship between regeneration and human demand is by tracking how much mutually exclusive, biologically productive area is necessary to provide the resource flows that people demand. Humans demand biologically productive areas in several quantifiable ways: production of food, fibre and timber; physical infrastructure such as roads and buildings; and absorption of waste, particularly the carbon dioxide from fossil fuel combustion. The total demand for biologically productive surfaces can be compared with the productive areas available that provide regeneration. Since the productivity of areas varies, they need to be measured not in terms of their physical extension, but in terms of biological regeneration they represent. For example, one can use a biologically productive hectare with world-average productivity as the common measurement unit that then allows expression of both demand and availability of productive areas in units that become comparable across space and time.Ecological footprint accounting is a well-documented concept to measure the total supply and demand of biological regeneration. In ecological footprint accounting, the ecosystem capacity to regenerate biomass is called biocapacity. It is measured in standardized ‘global hectares’, which represent the productivity of a world-average biologically productive hectare. The human demand for biocapacity is called the population’s ‘ecological footprint’, and it is the sum of all the mutually exclusive demands on these bioproductive areas. Ecological footprints are also expressed in global hectares.The principles of ecological footprint accounting, and the derived methods for national and sub-national assessments, are documented extensively within scientific literature6,7,8,9,38,44,45,46. The national accounting methodology has also been reviewed and documented by numerous national government agencies47.The essence of the approach is that regeneration is used as the lens to analyse both availability and demand because biological assets are materially the most limiting factor of the human economy1,2. In addition, biocapacity and ecological footprint can be tracked and compared with each other on the basis of two principles:

1.

By scaling every area proportionally to its biological productivity, each biologically productive area becomes commensurable with any other one. This is the essence of the global hectare.

2.

By including only areas that exclude other uses, that is, by making sure that every area is counted only once, the areas can meaningfully be added up, both for all the competing demands on productive surfaces (the ecological footprint) and for the surfaces that contain the planet’s regenerative capacity (the biocapacity).

The country-level accounts, called the National Footprint and Biocapacity Accounts, show that humanity’s demand exceeds Earth’s biocapacity, and the gap has been increasing since the 1970s8,38,40. This is consistent with research on planetary boundaries or ecosystem health1,2,10,11.Countries’ resource demand can be analysed from a consumption or a production perspective. The consumption perspective, which is the one used in this study, adjusts for trade and indicates the total resource consumption demand of a population. The production perspective identifies how much demand activities within a country directly put on ecosystems. This could be interpreted as the demand associated with generating the country’s GDP.Countries that demand more than their domestic ecosystems regenerate run a biocapacity deficit. It is made possible by three mechanisms: (1) overuse of domestic ecosystems, or local overshoot; (2) net import of biocapacity; and (3) use of the global commons, as in the case of emitting CO2 from fossil fuel into the atmosphere or fishing international waters38.Global results indicate that as of 2017, Earth had about 12.1 billion biologically productive hectares, according to Food and Agriculture Organization land-use statistics48. This includes productive ocean areas. By definition, this equals 12.1 billion global hectares, as each global hectare represents the productive average of all these 12.1 billion hectares. By contrast, human demand in 2017 added up to 20.9 billion global hectares, 73% higher than the regeneration of all the planet’s ecosystems combined (in per-person numbers, an average footprint of 2.8 global hectares contrasted to 1.6 global hectares of biocapacity available per person worldwide). This 73% overshoot may have dropped to 56% in 2020 due to lockdowns during COVID-194. In 2017, ecological footprint country averages varied from 0.5 global hectares per person (Eritrea) to 14.7 global hectares per person (Qatar). Biocapacity averages among countries stretch from 0.1 global hectares per person (Singapore) to 84 global hectares per person (Suriname)40.The accounts include only human demands (including domesticated animals) and not those of the millions of other living species, which together make possible the continuous functioning of the global ecosystem. To maintain biodiversity, which is critical for the integrity of the global ecosystem, humanity’s footprint would need to be less than the planet’s total biocapacity. E.O. Wilson, for example, proposed to only use half the planet’s capacity to secure 85% of its current biodiversity49. Using this objective as reference would imply that humanity’s current biological metabolism would be three times too large. It also makes clear that zero biocapacity deficits are a necessary but not sufficient condition for planetary resource stability. Still, for simplicity, we use the zero biocapacity deficit line as the demarcation line.Currently, the single-largest competing demand on the biosphere is the need for carbon sequestration capacity to neutralize emissions from fossil fuel burning. In 2020, this demand made up 57% of humanity’s ecological footprint. To comply with the Paris Agreement’s stated goal (Article 2 of ref. 30), this portion of the footprint would need to fall rapidly to zero. This reduction may come at the cost of increasing other parts of the ecological footprint. For example, more forest or agricultural products may be used to substitute for fossil fuels. If the Paris Agreement is fully implemented, there will be legal pressure to eliminate the carbon-related part of the deficit. If it is not implemented, the reduction pressures will emerge more slowly, which will increase the likelihood that the biocapacity will become increasingly damaged by climate change. Taking either path forces a country to eliminate its carbon footprint one way or the other. Fossil fuel dependence is therefore turning into an ever-growing risk. The pressure of increased land use has historically been the leading factor in the extinction of biodiversity, but unless nations can effectively control climate change, it will soon predominate as the major factor responsible for the massive extinction event that we humans have already started as a result of our unsustainable consumption—the sixth such event in the history of our planet.Measuring ability to purchase resources from abroadAnnual value production of an economy is measured by its GDP. It can be calculated as the value add of all its produced goods and services, as the sum of all the incomes or as the sum of all expenditures. Therefore, GDP can be used as a measure for a country’s income50,51.The analysis here focuses on the relative purchasing power of countries’ economic actors on global markets. Therefore, we use nominal US$ (or for time series, constant US$) instead of purchasing-power-adjusted US$, which reflect purchasing power on local markets. As economic actors compete for global resources in the same global market, each dollar has approximately the same weight, independent of the dollar’s purchasing power in the actor’s domestic market (called purchasing power parity). While this simplifies the fact that many commodities do not have a single homogeneous global market, the price range for resources in international markets is much narrower than that between domestic markets.For the sake of this analysis, we use average country income. Although incomes within countries vary vastly, we assume that nominal per-capita GDP is a reasonable approximation for national purchasing power in international markets. As a medium of exchange, money gives its owner the option to trade it in for physical assets, including biological resources; hence, more money means access to more resources.Not all international resource transfers are traded on global markets. Purchases could be under the protection of government-to-government arrangements or long-term contracts. The more of the international resource exchanges that occur in global markets, the tighter the competition on the global market for the remaining resources. Such increased competition makes the implications of the analysis presented here even more dramatic.In the context of global ecological overshoot, biocapacity scarcity will increase; therefore, the competition for purchasing additional resources will become even fiercer. In this case, using world-average income as an approximation for the dividing line between those who can net-purchase from abroad and those who cannot is too lenient. This demarcation indicates only that statistically those above the line can net-purchase from abroad. It does not indicate, however, whether they can purchase enough from abroad to cover their biocapacity deficit. This means that even more national economies than those identified by the 72% in this paper are excluded from being able to purchase sufficient resources from abroad. More

In the case of the first studied facility—Facility (I), thermo-visual examinations did not demonstrate any signs of thermal activity in the majority of the area. Above all, the new layered sections of the facility are free from thermal phenomena. However, the central section of the dump is characterized by rather intense thermal phenomena. The cone is also thermally active, yet in this section the thermal activity is not very intensive (see Fig. 1). In the section with the strongest thermal activity, i.e. the southern part of the top, the recorded surface temperature exceeded 600 °C (see Fig. 2). Outside the thermally active zone, the surface temperature did not exceed 25–35 °C.Figure 1Thermo-visual examination of the extractive waste dump (Facility I); October, 2017.Full size imageFigure 2Subsidence (the so-called crater) at the top of Facility I; October, 2017.Full size imageThe tests conducted on the premises of Facility II demonstrated the lack of thermal activity in the majority of the area. Only in the central section, on the top of the older part of the dump, the measurements showed the occurrence of thermal anomalies (see Fig. 3). These were minor areas located along the edge of the scarp on the northern and southern sides where the measured surface temperature was above 80 °C. It is worth mentioning that in 2010, the recorded temperatures for this section were at the level of several hundred degrees Celsius. Outside the sections, the temperature did not exceed 25–40 °C. The relatively high temperature in the area without thermal activity may be explained by the fact that in the morning hours it was exposed to the sun.Figure 3Thermo-visual examination of the extractive waste dump (Facility II); June, 2017.Full size imageDuring the course of the thermo-visual examination of the extractive waste dump in Facility III, no thermal activity was observed in the majority of the area. Only in the central section of the Facility, in the area which is still in operation, the measurements demonstrated the occurrence of thermal anomalies (see Fig. 4). These were minor areas where the local measured surface temperature was above 200 °C. Outside these areas, the temperature did not exceed 20–30 °C.Figure 4Thermo-visual examination of the extractive waste dump (Facility III); October, 2017.Full size imageThe simplest method of analyzing the differences and similarities between the studied objects is their visualization in the space of measured parameters. When two or three parameters are measured, such visualization is not problematic. However, in our study there are nine measured parameters. The graphic presentation of a 9-dimensional space is not possible. This is the rationale of applying the HCA which enables to analyze the similarities between the studied objects (three different dumping facilities in different periods of time) as well as the similarities between the parameters in object space. Nevertheless, the HCA does not allow to simultaneously analyze the relationships between the objects and the measured parameters. This problem was solved by the use of a color map of the experimental data, which enabled an in-depth interpretation of the data structure. In addition, the application of the color map facilitates highlighting the differences and similarities among the clusters showed in the dendrograms, and, in consequence, it helps to distinguish the facilities which are characterized by the highest or the lowest values of the measured parameters. Figure 5 demonstrates the dendrogram for 21 objects representing the studied dumping facilities in different periods of time in the space of 9 measured parameters, the dendrogram for the measured parameters in the object space as well as a color map presenting the values of the measured parameters for particular objects.Figure 5Dendrograms for (a) 21 objects representing the studied dumping facilities (see Table 1) in the space of 9 measured parameters; (b) parameters in the object space; (c) a color map presenting the values of measured parameters for particular dumping facilities.Full size imageBased on the dendrogram presented in Fig. 5a, a clear distinction of the examined objects representing the discussed dumping facilities into two clusters—A and B can be observed. Cluster A includes all samples representing Facility II in the whole period of the monitoring as well as samples representing Facility I taken in Quarter 3, 2017 and in Quarters 1–4, 2018 (objects nos. 1 and 3–6). All samples representing Facility III as well as two samples representing Facility I taken in Quarter 4, 2017 and in Quarter 4, 2018, respectively (objects nos. 2 and 6) are collected in Cluster B. In addition, within each of the clusters certain sub-clusters may be distinguished. Within Cluster A, the following sub-groups can be observed:

sub-group A1 collecting all samples taken from Facility II during the whole period of the monitoring (objects nos. 8–14);

sub-group A2 including samples taken from Facility I in Quarter III, 2017 and in Quarters 1–4, 2018 (objects nos. 1 and 3–6).

In turn, Cluster B contains the following three sub-groups:

sub-group B1 collecting two samples taken from Facility I in Quarter 4, 2017 and 2018, respectively (objects nos. 2 and 6) as well as samples representing Facility III in Quarter 3, 2017, Quarters 1–3, 2018 and Quarter 1, 2019 (objects nos. 15, 17–19 and 21);

sub-group B2 encompassing the remaining two samples taken from Facility III in Quarter 4, 2017 and in Quarter 4, 2018 (objects nos. 16 and 20).

The dendrogram obtained by means of Ward’s linkage method for the measured parameters in the space of 21 objects (see Fig. 5b) enables to distinguish the three principal clusters of parameters listed below:

class A collecting parameters nos. 2, 3 and 4 (describing the concentrations of acenaphtene, fluorene and phenanthrene);

class B containing parameters nos. 1 and 9 (describing the concentrations of naphthalene and chrysene);

class C including the remaining parameters nos. 5, 6, 7 and 8 (describing the concentrations of anthracene, pyrene, fluoranthene and B(a)anthracene).

The PAHs emissions from a burning mine waste dump must be carefully monitored due to their potential toxicity and genotoxicity38,39. PAHs have two different roots; one is incomplete combustion of organic matter, whereas the other one is their production in the geological formation when organic sediments were chemically transformed into fossil fuels. In our study, the first path, namely spontaneous coal waste combustion is observed40,41. It is also worth mentioning that PAHs pose significant human health hazards. The exposure to PAHs may result in skin, lung or stomach cancers in the human organism28.As mentioned before, a considerable drawback of the Hierarchical Clustering Analysis lies in that it does not allow for simultaneous interpretation of the dendrograms describing objects which represent particular samples taken from the extractive waste facility in different periods in the space of measured parameters and parameters in the object space. The lack of the possibility of interpreting the above relationships significantly limits the knowledge of the studied phenomena because the purpose of the analysis is to determine not only the differences among particular samples but also the underlying cause of such differences. Therefore, the dendrogram presenting the examined samples taken from the three selected facilities in different periods of time (see Fig. 5a) was juxtaposed with a color map of the experimental data (see Fig. 5c) demonstrating the values of the measured parameters arranged according to the order of the objects and parameters organization. The juxtaposition enables to determine the reason why the examined samples were distributed in such a way. In addition, the interpretation of the dendrogram for the objects in parameters space complemented with the color map of experimental data allows distinguishing samples which are characterized by the highest values of the measured parameters.Analyzing the dendrogram presented in Fig. 5a together with the color map of the experimental data, it can be observed that all samples within Cluster A were characterized by relatively lower values of the measured parameters. Moreover, sub-group A1 identified within Cluster A was characterized by relatively lowest concentrations of acenaphtene, fluorene, phenanthrene and pyrene (parameters nos. 2, 3, 4, and 7) among all of the examined samples taken from the three selected facilities in the whole period of the monitoring. It confirms the observations which were made previously with the use of a thermo-visual camera which demonstrated that there were no thermal phenomena in the majority of the waste dump areas. The observed sparse emissions of gases result from certain thermal anomalies occurring in the central section of Facility II as well as some small areas located along the northern and southern side of the scarp; however, in any of the places the temperature did not exceed 80 °C (see Fig. 3). The thermo-visual examinations of the remaining sections of the facility showed that the surface temperature did not exceed the level of 25–40 °C and it is a result of sun exposure rather than the occurrence of any thermal phenomena.Sub-group A2 which includes all samples taken from Facility I in Quarter 3, 2017 and Quarters 1–4, 2018 (objects nos. 1 and 3–6) differs from A1 objects in terms of relatively higher concentrations of acenaphtene, fluorene, phenanthrene and pyrene (parameters nos. 2, 3, 4, and 7). Additionally, within sub-group A2, the uniqueness of sample 1 taken in Quarter 3, 2018 (object no. 5) can be observed; it was characterized by the lowest concentration of chryzene (parameter no. 9) of all the examined samples.Despite the fact that the results of the thermo-visual analysis of Facility III in general did not show any major signs of thermal activity, the Hierarchical Clustering Analysis by means of which the samples were examined in terms of the emission of the 9 parameters indicated that some thermal phenomena take place at this dump.Furthermore, the analysis of the color map of the experimental data for the objects grouped in Cluster B makes it abundantly clear that thermal activity takes place in the majority of the monitored areas located at Facility III. What is more, for two samples taken from Facility I in Quarter 4, 2017 and Quarter 4, 2018 (objects nos. 2 and 6), thermal activity was observed. For the two objects, the concentrations of acenaphtene and fluorene are the highest among all of the examined samples (parameters nos. 2 and 3); similarly, the concentrations of phenanthrene, pyrene and B(a)anthracene (parameters nos. 4, 7, and 8) are also high. It confirms the observations made by means of a thermo-visual camera which demonstrated that there were no thermal phenomena if one considers the facility in its entirety. However, there are certain areas in the central section of the facility where thermal phenomena occur.In the case of the samples taken from Facility III in Quarter 3, 2017 as well as in Quarter 2 and Quarter 3, 2018 (objects nos. 15, 18, and 19) within sub-group B1, the values of all measured parameters are only slightly increased, which can indicate the occurrence of thermal phenomena. Yet, it must be kept in mind that the concentrations are decidedly lower than for the remaining samples taken from Facility III. It may be also an indication that self-heating of the coal extractive waste takes place. The other two samples taken from Facility III in Quarter 1, 2018 and Quarter 1, 2019 (objects nos. 17 and 21) are characterized by relatively highest concentrations of pyrene and B(a)anthracene (parameters nos. 7 and 8) as well as high concentrations of anthracene and fluorene (parameters nos. 5 and 6), which can result from an intense thermal activity.A similar observation can be made for the two samples taken from Facility III in Quarter 4, 2017 and Quarter 4, 2018 (objects nos. 16 and 20) which are characterized by the highest concentrations of naphthalene, anthracene, fluoranthene and chrysene of all the examined samples (parameters nos. 1, 5, 6 and 9) as well as high concentrations of phenanthrene (parameter no. 4). Although the results of the thermo-visual analysis of the whole area of Facility III did not demonstrate signs of thermal activity in most of its sections, there exist certain spots with intense thermal processes, which is confirmed by relatively high values of all the measured parameters. More