Ecology

1.
Serre, D. & Paabo, S. Evidence for gradients of human genetic diversity within and among continents. Genome Res. 14, 1679–1685 (2004).
CAS  PubMed  PubMed Central  Article  Google Scholar 
2.
Ohta, T. Population size and rate of evolution. J. Mol. Evol. 1, 305–314 (1972).
PubMed  Article  PubMed Central  Google Scholar 

3.
Slatkin, M. Gene flow and selection in a cline. Genetics 75, 733–756 (1973).

4.
Potts, R. Variability selection in hominid evolution. Evol. Anthropol. 7, 81–96 (1998).
Article  Google Scholar 

5.
Potts, R. Hominin evolution in settings of strong environmental variability. Quat. Sci. Rev. 73, 1–13 (2013).
Article  Google Scholar 

6.
Prüfer, K. et al. The bonobo genome compared with the chimpanzee and human genomes. Nature 486, 527–531 (2012).
PubMed  PubMed Central  Article  CAS  Google Scholar 

7.
Hill, W. C. O. in The Nomenclature, Taxonomy and Distribution of Chimpanzees, Vol. 1 (ed. Bourne, G. H.) 22–49 (Karger, 1969).

8.
Pruetz, J. D. & Bertolani, P. Chimpanzee (Pan troglodytes verus) behavioral responses to stresses associated with living in a savannah-mosaic environment: implications for hominin adaptations to open habitats. Paleoanthropology 2009, 252–262 (2009).
Article  Google Scholar 

9.
Fünfstück, T. et al. The sampling scheme matters: Pan troglodytes troglodytes and P. t. schweinfurthii are characterized by clinal genetic variation rather than a strong subspecies break. Am. J. Phys. Anthropol. 156, 181–191 (2014).
PubMed  PubMed Central  Article  Google Scholar 

10.
Fischer, A., Pollack, J., Thalmann, O., Nickel, B. & Paabo, S. Demographic history and genetic differentiation in apes. Curr. Biol. 16, 1133–1138 (2006).
CAS  PubMed  Article  Google Scholar 

11.
Gonder, M. K. et al. A new west African chimpanzee subspecies? Nature 388, 337 (1997).
CAS  PubMed  Article  Google Scholar 

12.
Becquet, C., Patterson, N., Stone, A. C., Przeworski, M. & Reich, D. E. Genetic structure of chimpanzee populations. PLoS Genet. 3, e66 (2007).
PubMed  PubMed Central  Article  CAS  Google Scholar 

13.
Bowden, R. et al. Genomic tools for evolution and conservation in the chimpanzee: Pan troglodytes ellioti is a genetically distinct population. PLoS Genet. 8, e1002504 (2012).
CAS  PubMed  PubMed Central  Article  Google Scholar 

14.
Prado-Martinez, J. et al. Great ape genetic diversity and population history. Nature 499, 471–475 (2013).
CAS  PubMed  PubMed Central  Article  Google Scholar 

15.
de Manuel, M. et al. Chimpanzee genomic diversity reveals ancient admixture with bonobos. Science 354, 477–481 (2016).
PubMed  PubMed Central  Article  CAS  Google Scholar 

16.
Langergraber, K. E. et al. Genetic and ‘cultural’ similarity in wild chimpanzees. Proc. R. Soc. B. 278, 408–416 (2010).
PubMed  Article  PubMed Central  Google Scholar 

17.
Garcin, Y. et al. Early anthropogenic impact on Western Central African rainforests 2,600 y ago. Proc. Natl Acad. Sci. USA 115, 3261–3266 (2018).
CAS  PubMed  Article  PubMed Central  Google Scholar 

18.
Vicente, M. & Schlebusch, C. M. African population history: an ancient DNA perspective. Curr. Opin. Genet. Dev. 62, 8–15 (2020).
CAS  PubMed  Article  PubMed Central  Google Scholar 

19.
IUCN Red List. IUCN red list of threatened species. http://www.iucnredlist.org (2020).

20.
Wentworth, C. K. Natural bridges and glaciation. Am. J. Sci. 26, 577–584 (1933).
Article  Google Scholar 

21.
Maley, J. The African rainforest: main characteristics of changes in vegetation and climate from the Upper Cretaceous to the Quaternary. Proc. R. Soc. Edinb. 104B, 31–73 (1996).
Google Scholar 

22.
Langergraber, K. E. et al. Generation times in wild chimpanzees and gorillas suggest earlier divergence times in great ape and human evolution. Proc. Natl Acad. Sci. USA 109, 15716–15721 (2012).
CAS  PubMed  Article  Google Scholar 

23.
Shea, B. T. & Coolidge, H. J. Craniometric differentiation and systematics in the genus Pan. J. Hum. Evol. 13, 671–685 (1988).
Article  Google Scholar 

24.
Albrecht, G. H. & Miller, J. M. A. in Geographic Variation in Primates (eds Kimbel, W. H. & Martin, L. B.) 123–161 (Springer Science & Business Media, 2013).

25.
Shea, B. T., Leigh, S. R. & Groves, C. P. in Multivariate Craniometric Variation in Chimpanzees (eds Kimbel, W. H. & Martin, L. B.) 265–296 (Springer Science & Business Media, 2013).

26.
Kühl, H. S. et al. Chimpanzee accumulative stonethrowing. Sci. Rep. 6, 22219 (2016).

27.
Wright, S. Isolation by distance. Genetics 28, 114–138 (1943).
CAS  PubMed  PubMed Central  Google Scholar 

28.
Pritchard, J. K., Stephens, M. & Donnelly, P. Inference of population structure using multilocus genotype data. Genetics 155, 945–959 (2000).
CAS  PubMed  PubMed Central  Google Scholar 

29.
Pritchard, J. K., Wen, X. & Falush, D. Documentation for structure version 2.3 software: version 2.3. 1–39. http://pritchardlab.stanford.edu/structure_software/release_versions/v2.3.4/structure_doc.pdf (2009).

30.
Meirmans, P. G. The trouble with isolation by distance. Mol. Ecol. 21, 2839–2846 (2012).
PubMed  Article  Google Scholar 

31.
Perez, M. F. et al. Assessing population structure in the face of isolation by distance: are we neglecting the problem? Divers. Distrib. 24, 1883–1889 (2018).
Article  Google Scholar 

32.
Thalib, L., Kitching, R. L. & Bhatti, M. I. Principal component analysis for grouped data – a case study. Environmetrics 10, 565–574 (1999).
Article  Google Scholar 

33.
Fischer, A. et al. Bonobos fall within the genomic variation of chimpanzees. PLoS ONE 6, e21605 (2011).
CAS  PubMed  PubMed Central  Article  Google Scholar 

34.
Frantz, A. C., Cellina, S., Krier, A., Schley, L. & Burke, T. Using spatial Bayesian methods to determine the genetic structure of a continuously distributed population: clusters or isolation by distance? J. Appl. Ecol. 46, 493–505 (2009).
Article  Google Scholar 

35.
Kalinowski, S. T. The computer program STRUCTURE does not reliably identify the main genetic clusters within species: simulations and implications for human population structure. Heredity 106, 625–632 (2010).
PubMed  PubMed Central  Article  Google Scholar 

36.
Schwartz, M. K. & McKelvey, K. S. Why sampling scheme matters: the effect of sampling scheme on landscape genetic results. Conserv. Genet. 10, 441–452 (2008).
Article  Google Scholar 

37.
Meirmans, P. G. Seven common mistakes in population genetics and how to avoid them. Mol. Ecol. 24, 3223–3231 (2015).
PubMed  Article  PubMed Central  Google Scholar 

38.
Arandjelovic M, et al. Pan African Programme—The cultured chimpanzee. Guidelines for research and data collection. http://panafrican.eva.mpg.de/english/approaches_and_methods.php (2014).

39.
Arandjelovic, M. & Vigilant, L. Non-invasive genetic censusing and monitoring of primate populations. Am. J. Primatol. 80, e22743 (2018).
PubMed  Article  Google Scholar 

40.
Arthofer, W., Heussler, C., Krapf, P., Schlick-Steiner, B. C. & Steiner, F. M. Identifying the minimum number of microsatellite loci needed to assess population genetic structure: a case study in fly culturing. Fly 12, 13–22 (2018).

41.
Wenburg, J. K., Bentzen, P. & Foote, C. J. Microsatellite analysis of genetic population structure in an endangered salmonid: the coastal cutthroat trout (Oncorhyncus clarki clarki). Mol. Ecol. 7, 733–749 (1998).
CAS  Article  Google Scholar 

42.
Guo, X.-Z. et al. Phylogeography and populationgenetics of Schizothorax o’connori: strong subdivision in the Yarlung Tsangpo River inferred from mtDNA and microsatellite markers. Sci. Rep. 6, 29821 (2016).

43.
Kleinhans, C. & Willows-Munro, S. Low genetic diversity andshallow population structure inthe endangered vulture, Gyps copotheres. Sci. Rep. 9, 5536 (2019).
PubMed  PubMed Central  Article  CAS  Google Scholar 

44.
Bonato, L. et al. Diversity among peripheral populations: genetic and evolutionary differentiation of Salamandra atraat the southern edge of the Alps. J. Zool. Syst. Evol. Res. 56, 533–548 (2018).
Article  Google Scholar 

45.
Balkenhol, N. et al. A multi-method approach for analyzing hierarchical genetic structures: a case study with cougars Puma concolor. Ecography 37, 552–563 (2014).
Article  Google Scholar 

46.
Kobayashi, T. & Sota, T. Contrasting effects of habitat discontinuity on three closely related fungivorous beetle species with diverging host‐use patterns and dispersal ability. Ecol. Evol. 9, 2475–2486 (2019).
PubMed  PubMed Central  Article  Google Scholar 

47.
Rousset, F. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145, 1219–1228 (1997).

48.
Valdes, A. M., Slatkin, M. & Freimer, N. B. Allele frequencies at microsatellite loci: the stepwise mutation model revisited. Genetics 133, 737–749 (1993).
CAS  PubMed  PubMed Central  Article  Google Scholar 

49.
Goldstein, D. B., Ruiz Linares, A., Cavalli-Sforza, L. L. & Feldman, M. W. An evaluation of genetic distances for use with microsatellite loci. Genetics 139, 463–471 (1995).
CAS  PubMed  PubMed Central  Article  Google Scholar 

50.
Calabrese, P. P., Durrett, R. T. & Aquadro, C. F. Dynamics of microsatellite divergence under stepwise mutation and proportional slippage/point mutation models. Genetics 159, 839–852 (2001).
CAS  PubMed  PubMed Central  Google Scholar 

51.
Petkova, D., Novembre, J. & Stephens, M. Visualizing spatial population structure with estimated effective migration surfaces. Nat. Genet. 48, 94–100 (2015).
PubMed  PubMed Central  Article  CAS  Google Scholar 

52.
Rich, A. M., Wasserman, M. D., Hunt, K. D. & Kaestle, F. A. Chimpanzee (Pan troglodytes schweinfurthii) population spans multiple protected areas in the Albertine Rift. Folia. Primatol. 91, 595–609 (2020).

53.
Baldwin, P. J., McGrew, W. C. & Tutin, C. E. G. Wide-ranging chimpanzees at Mt. Assirik, Senegal. Int. J. Primatol. 3, 367–385 (1982).
Article  Google Scholar 

54.
Lemoine, S. et al. Group dominance increases territory size and reduces neighbour pressure in wild chimpanzees. R. Soc. Open Sci. 7, 200577 (2020).
PubMed  PubMed Central  Article  Google Scholar 

55.
Newton-Fisher, N. E. The home range of the Sonso community of chimpanzees from the Budongo Forest, Uganda. Afr. J. Ecol. 41, 150–156 (2003).
Article  Google Scholar 

56.
Allendorf, F. W. Genetic drift and the loss of alleles versus heterozygosity. Zoo. Biol. 5, 181–190 (1986).
Article  Google Scholar 

57.
Barratt, C. D., et al. Late Quaternary habitat suitability models for chimpanzees (Pan troglodytes) since the Last Interglacial (120,000 BP). Preprint at BioRxiv https://www.biorxiv.org/content/10.1101/2020.05.15.066662v1 (2019).

58.
Bertola, L. D. et al. Phylogeographic patterns in Africa and high resolution delineation of genetic clades in the lion (Panthera leo). Sci. Rep. 6, 30807 (2016).

59.
Marchesi, P., Marchesi, N., Fruth, B. & Boesch, C. Census and distribution of chimpanzees in Côte D’Ivoire. Primates 36, 591–607 (1995).
Article  Google Scholar 

60.
Sommer, V., Adanu, J., Faucher, I. & Fowler, A. Nigerian chimpanzees (Pan troglodytes vellerosus) at Gashaka: two years of habituation efforts. Folia Primatol. 75, 295–316 (2004).
Article  Google Scholar 

61.
Chancellor, R. L., Langergraber, K. E., Ramirez, S., Rundus, A. S. & Vigilant, L. Genetic sampling of unhabituated chimpanzees (Pan troglodytes schweinfurthii) in Gishwati Forest Reserve, an isolated forest fragment in western Rwanda. Int. J. Primatol. 33, 479–488 (2012).
Article  Google Scholar 

62.
Piel, A. K. et al. Population status of chimpanzees in the Masito-Ugalla Ecosystem, Tanzania. Am. J. Primatol. 77, 1027–1035 (2015).
PubMed  Article  Google Scholar 

63.
Wessling, E. G. et al. Seasonal variation in physiology challenges the notion of chimpanzees as a forest-adapted species. Front. Ecol. Evol. 6, 1–21 (2018).
Article  Google Scholar 

64.
Whiten, A. et al. Cultures in chimpanzees. Nature 399, 682–685 (1999).
CAS  PubMed  Article  Google Scholar 

65.
Kühl, H. S. et al. Human impact erodes chimpanzee behavioral diversity. Science 363, 1453–1455 (2019).
PubMed  Article  CAS  Google Scholar 

66.
Luncz, L. V., Mundry, R. & Boesch, C. Evidence for cultural differences between neighboring chimpanzee communities. Curr. Biol. 22, 922–926 (2012).
CAS  PubMed  Article  Google Scholar 

67.
Boesch, C., Marchesi, P., Marchesi, N., Fruth, B. & Joulian, F. Is nut cracking in wild chimpanzees a cultural behaviour? J. Hum. Evol. 26, 325–338 (1994).
Article  Google Scholar 

68.
Kalan, A. K. et al. Environmental variability supports chimpanzee behavioural diversity. Nat. Commun. 11, 4451 (2020).

69.
Hockings, K. J. in Chimpanzees of Bossou and Nimba, 221–219 (Springer Science & Business Media, 2011).

70.
McCarthy, M. S., Lester, J. D. & Stanford, C. B. Chimpanzees (Pan troglodytes) flexibly use introduced species for nesting and bark feeding in a human-dominated habitat. Int. J. Primatol. 38, 321–337 (2016).
PubMed  PubMed Central  Article  Google Scholar 

71.
McLennan, M. R. et al. Surviving at the extreme: chimpanzee ranging is not restricted in a deforested human‐dominated landscape in Uganda. Afr. J. Ecol. 8, e57872 (2020).
Google Scholar 

72.
Junker, J. et al. Recent decline in suitable environmental conditions for African great apes. Divers. Distrib. 18, 1077–1091 (2012).
Article  Google Scholar 

73.
Kühl, H. S. et al. The Critically Endangered western chimpanzee declines by 80%. Am. J. Primatol. 79, e22681 (2017).
Article  Google Scholar 

74.
Walsh, P. D., Breuer, T., Sanz, C., Morgan, D. B. & Doran-Sheehy, D. M. Potential for Ebola transmission between gorilla and chimpanzee social groups. Am. Nat. 169, 684–689 (2007).
PubMed  Article  Google Scholar 

75.
Baden, A. L. et al. Anthropogenic pressures drive population genetic structuring across a critically endangered lemur species range. Sci. Rep. 9, 16276 (2019).
PubMed  PubMed Central  Article  CAS  Google Scholar 

76.
Nsubuga, A. M. et al. Factors affecting the amount of genomic DNA extracted from ape faeces and the identification of an improved sample storage method. Mol. Ecol. 13, 2089–2094 (2004).
CAS  PubMed  Article  Google Scholar 

77.
Arandjelovic, M. et al. Two-step multiplex polymerase chain reaction improves the speed and accuracy of genotyping using DNA from noninvasive and museum samples. Mol. Ecol. Resour. 9, 28–36 (2009).
CAS  PubMed  Article  Google Scholar 

78.
Kalinowski, S. T., Taper, M. L. & Marshall, T. C. Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment. Mol. Ecol. 16, 1099–1106 (2007).
PubMed  Article  Google Scholar 

79.
Waits, L. P., Luikart, G. & Taberlet, P. Estimating the probability of identity among genotypes in natural populations: cautions and guidelines. Mol. Ecol. 10, 249–256 (2001).
CAS  PubMed  Article  Google Scholar 

80.
Meirmans, P. G. & Hedrick, P. W. Assessing population structure: FST and related measures. Mol. Ecol. Resour. 11, 5–18 (2010).
PubMed  Article  PubMed Central  Google Scholar 

81.
Mondol, S. et al. New evidence for hybrid zones of forest and savanna elephants in Central and West Africa. Mol. Ecol. 24, 6134–6147 (2015).
CAS  PubMed  Article  PubMed Central  Google Scholar  More

1.
Bindoff, N. L. et al. Changing Ocean, Marine Ecosystems, and Dependent Communities. in IPCC Special Report on the Ocean and Cryosphere in a Changing Climate (eds. Pörtner, H.-O. et al.) 447–588 (IPCC, 2019).
2.
Collins, M. et al. Extremes, Abrupt Changes and Managing Risks. in IPCC Special Report on the Ocean and Cryosphere in a Changing Climate (eds. Pörtner, H.-O. et al.) 589–656 (IPCC, 2019).

3.
Hoegh-Guldberg et al. The Ocean. in Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds. Barros, V. R. et al.) 1655–1731 (Cambridge University Press, 2014).

4.
Laufkotter, C., Zscheischler, J. & Frolicher, T. L. High-impact marine heatwaves attributable to human-induced global warming. Science 369, 1621–1625 (2020).
CAS  PubMed  PubMed Central  Google Scholar 

5.
Holbrook, N. J. et al. A global assessment of marine heatwaves and their drivers. Nat. Commun. 10, 1–13 (2019).
CAS  Article  Google Scholar 

6.
Oliver, E. C. J. et al. Longer and more frequent marine heatwaves over the past century. Nat. Commun. 9, 1–12 (2018).
CAS  Article  Google Scholar 

7.
Carr, M. & Kearns, E. J. Production regimes in four Eastern Boundary Current systems. Deep. Res. Part II 50, 3199–3221 (2003).
CAS  Article  Google Scholar 

8.
Castro, C. G. et al. Introduction to ‘ The 1997 – 8 El Nino Atlas of oceanographic conditions along the west coast of North America (23°N – 50°N)’. Prog. Oceanogr. 54, 503–511 (2002).
Article  Google Scholar 

9.
Kendrick, G. A. et al. A systematic review of how multiple stressors from an extreme event drove ecosystem-wide loss of resilience in an iconic seagrass community. Front. Mar. Sci. 6, 1–15 (2019).
Article  Google Scholar 

10.
Nohaïc, M. L. et al. Marine heatwave causes unprecedented regional mass bleaching of thermally resistant corals in northwestern Australia. Sci. Rep. 7, 1–11 (2017).
Article  CAS  Google Scholar 

11.
Smale, D. A. Impacts of ocean warming on kelp forest ecosystems. N. Phytol. 225, 1447–1454 (2020).
Article  Google Scholar 

12.
Smale, D. A. et al. Marine heatwaves threaten global biodiversity and the provision of ecosystem services. Nat. Clim. Chang. 9, 306–312 (2019).
Article  Google Scholar 

13.
Schiel, D. R. & Foster, M. S. The Biology and Ecology of Giant Kelp Forests (University of California Press, 2015).

14.
Wernberg, T., Krumhansl, K., Filbee-dexter, K. & Pedersen, M. F. in World Seas: An Environmental Evaluation (ed. Sheppard, C.) 57–78 (Elsevier Ltd., 2019).

15.
Wernberg, T. et al. An extreme climatic event alters marine ecosystem structure in a global biodiversity hotspot. Nat. Clim. Chang. 3, 78–82 (2012).
Article  Google Scholar 

16.
Wernberg, T. et al. Climate-driven regime shift of a temperate marine ecosystem. Science 353, 169–172 (2016).
CAS  PubMed  PubMed Central  Article  Google Scholar 

17.
Oliver, E. C. J. et al. The unprecedented 2015/16 Tasman Sea marine heatwave. Nat. Commun. 8, 1–12 (2017).
Article  Google Scholar 

18.
Thomsen, M. S. et al. Local extinction of bull kelp (Durvillaea spp.) due to a marine heatwave. Front. Mar. Sci. 6, 1–10 (2019).
Article  Google Scholar 

19.
Cavanaugh, K. C. et al. Spatial variability in the resistance and resilience of giant kelp in Southern and Baja California to a multiyear heatwave. Front. Mar. Sci. 6, 1–14 (2019).
Article  Google Scholar 

20.
Arafeh-dalmau, N., Montaño-moctezuma, G., Martínez, J. A. & Smale, D. A. Extreme marine heatwaves alter kelp forest community near its equatorward distribution limit. Front. Mar. Sci. 6, 1–18 (2019).
Article  Google Scholar 

21.
Filbee-Dexter, K., Feehan, C. J. & Scheibling, R. E. Large-scale degradation of a kelp ecosystem in an ocean warming hotspot. Mar. Ecol. Prog. Ser. 543, 141–152 (2016).

22.
Rogers-Bennett, L. & Catton, C. A. Marine heat wave and multiple stressors tip bull kelp forest to sea urchin barrens. Sci. Rep. 9, 1–9 (2019).
CAS  Article  Google Scholar 

23.
Sanford, E., Sones, J. L., García-reyes, M., Goddard, J. H. R. & Largier, J. L. Widespread shifts in the coastal biota of northern California during the 2014- 2016 marine heatwaves. Sci. Reportscientific Rep. 9, 1–14 (2019).
Article  CAS  Google Scholar 

24.
Beas-Luna, R. et al. Geographic variation in responses of kelp forest communities of the California current to recent climatic changes. Glob. Chang. Biol. https://doi.org/10.1111/gcb.15273 (2020).

25.
Filbee-Dexter, K. & Wernberg, T. Rise of turfs: a new battlefront for globally declining kelp forests. Bioscience 68, 64–76 (2018).
Article  Google Scholar 

26.
Filbee-Dexter, K. & Scheibling, R. E. Sea urchin barrens as alternative stable states of collapsed kelp ecosystems. Mar. Ecol. Prog. Ser. 495, 1–25 (2014).
Article  Google Scholar 

27.
Ling, S. D., Johnson, C. R., Frusher, S. D. & Ridgway, K. R. Overfishing reduces resilience of kelp beds to climate-driven catastrophic phase shift. Proc. Natl. Acad. Sci. USA 106, 22341–22345 (2009).

28.
Estes, J. A., Tinker, M. T., Williams, T. M. & Doak, D. F. Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science 282, 473–477 (1998).
CAS  PubMed  Article  PubMed Central  Google Scholar 

29.
Hamilton, S. L. & Caselle, J. E. Exploitation and recovery of a sea urchin predator has implications for the resilience of southern California kelp forests. Proc. R. Soc. B 282, 1–10 (2014).

30.
Jacox, M. G. et al. Forcing of multiyear extreme ocean temperatures that impacted California current living marine resources in 2016. Bull. Am. Meteorol. Soc. 99, 27–33 (2018).
Article  Google Scholar 

31.
Montecino-Latorre, D. et al. Devastating transboundary impacts of sea star wasting disease on subtidal asteroids. PLoS ONE 11, 1–21 (2016).
Article  CAS  Google Scholar 

32.
Eisaguirre, J. M., Davis, K., Carlson, P. M., Gaines, S. D. & Caselle, J. E. Trophic redundancy and predator size class structure drive differences in kelp forest ecosystem dynamics. Ecology 101, 1–11 (2020).
Article  Google Scholar 

33.
Harrold, C. & Reed, D. C. Food availability, sea urchin grazing, and kelp forest community structure. Ecology 66, 1160–1169 (1985).
Article  Google Scholar 

34.
Cowen, R. K. The effect of sheephead (Semicossyphus pulcher) predation on red sea urchin (Strongylocentrotus franciscanus) populations: an experimental analysis. Oecologia 58, 249–255 (1983).
PubMed  Article  Google Scholar 

35.
Burt, J. M. et al. Sudden collapse of a mesopredator reveals its complementary role in mediating rocky reef regime shifts. Proc. R. Soc. B 285, 1–9 (2018).

36.
Springer, Y. P., Hays, C., Carr, M. H. & Mackey, M. R. Toward ecosystem-based management of marine macroalgae-the bull kelp, Nereocystis luetkeana. Oceanogr. Mar. Biol. Annu. Rev. 48, 1–42 (2010).
Google Scholar 

37.
Springer, Y., Hays, C., Carr, M., Mackey, M. & Bloeser, J. Ecology and management of the bull kelp, Nereocystis luetkeana: a synthesis with recommendations for future research. Vol. 48, 1–45 (Lenfest Ocean Program at The Pew Charitable Trusts, 2006).

38.
Bell, T. W., Allen, J. G., Cavanaugh, K. C. & Siegel, D. A. Three decades of variability in California’s giant kelp forests from the Landsat satellites. Remote Sens. Environ. 238, 110811 (2020).

39.
Pfister, C. A., Berry, H. D. & Mumford, T. The dynamics of Kelp Forests in the Northeast Pacific Ocean and the relationship with environmental drivers. J. Ecol. https://doi.org/10.1111/1365-2745.12908 (2017).

40.
Griggs, G. B. & Hein, J. R. Sources, dispersal, and clay mineral composition of fine-grained sediment off Central and Northern California. J. Geol. 88, 541–566 (1979).
Article  Google Scholar 

41.
Harvell, C. D., Caldwell, J. M., Burt, J. M., Bosley, K. & Keller, A. Disease epidemic and a marine heat wave are associated with the continental-scale collapse of a pivotal predator (Pycnopodia helianthoides). Sci. Adv. 5, 1–9 (2019).
Article  Google Scholar 

42.
Okamoto, D. K., Schroeter, S. C. & Reed, D. C. Effects of ocean climate on spatiotemporal variation in sea urchin settlement and recruitment. Limnol. Oceanogr. https://doi.org/10.1002/lno.11440 (2020).

43.
Estes, J. A., Burdin, A. & Doak, D. F. Sea otters, kelp forests, and the extinction of Steller’s sea cow. Proc. Natl Acad. Sci. USA 113, 880–885 (2016).
CAS  PubMed  Article  PubMed Central  Google Scholar 

44.
Scheffer, M., Carpenter, S., Foley, J. A., Folke, C. & Walker, B. Catastrophic shifts in ecosystems. Nature 413, 591–596 (2001).
CAS  PubMed  Article  PubMed Central  Google Scholar 

45.
Ebert, T. A., Schroeter, S. C., Dixon, J. D. & Kalvass, P. Settlement patterns of red and purple sea urchins (Strongylocentrotus franciscanus and S. purpuratus) in California, USA. Mar. Ecol. Prog. Ser. 111, 41–52 (1994).
Article  Google Scholar 

46.
Frölicher, T. L., Fischer, E. M. & Gruber, N. Marine heatwaves under global warming. Nature 560, 360–366 (2018).
PubMed  Article  CAS  PubMed Central  Google Scholar 

47.
Steneck, R. S. et al. Kelp forest ecosystems: biodiversity, stability, resilience and future. Environ. Conserv. 29, 436–459 (2002).
Article  Google Scholar 

48.
Smith, J. G. et al. Behavioral responses across a mosaic of ecosystem states restructure a sea otter-urchin trophic cascade. Proc. Natl. Acad. Sci. USA 118, 202012493 (2020). https://doi.org/10.1073/pnas.2012493118.

49.
Eurich, J. G., Selden, R. L. & Warner, R. R. California spiny lobster preference for urchins from kelp forests: implications for urchin barren persistence. Mar. Ecol. Prog. Ser. 498, 217–225 (2014).
Article  Google Scholar 

50.
Graham, M. H. Effects of local deforestation on the diversity and structure of Southern California giant kelp forest food webs. Ecosystems 7, 341–357 (2004).
Article  Google Scholar 

51.
Babcock, R. C. et al. Decadal trends in marine reserves reveal differential rates of change in direct and indirect effects. Proc. Natl Acad. Sci. USA 107, 18256–18261 (2010).
CAS  PubMed  Article  PubMed Central  Google Scholar 

52.
Hewson, I., Bistolas, K. S. I., Cardé, E. M. Q. & Button, J. B. Investigating the complex association between viral ecology, environment, and Northeast Pacific Sea Star Wasting. Front. Mar. Sci. 5, 1–14 (2018).

53.
Pearse, J. S. & Hines, A. H. Expansion of a central California kelp forest following the mass mortality of sea urchins. Mar. Biol. 51, 83–91 (1979).

54.
Ebeling, A. W., Laur, D. R., Rowley, R. J. & Barbara, S. Severe storm disturbances and reversal of community structure in a southern California kelp forest. Mar. Biol. 294, 287–294 (1985).
Article  Google Scholar 

55.
Martínez, B. et al. Distribution models predict large contractions of forming seaweeds in response to ocean warming. Divers. Distrib. 24, 1350–1366 (2018).
Article  Google Scholar 

56.
DeYoung, B. et al. Regime shifts in marine ecosystems: detection, prediction and management. Trends Ecol. Evol. 23, 402–409 (2008).
PubMed  Article  Google Scholar 

57.
Crépin, A., Biggs, R., Polasky, S., Troell, M. & Zeeuw, A. De. Regime shifts and management. Ecol. Econ. 84, 15–22 (2012).
Article  Google Scholar 

58.
Roberts, D. A. et al. Mapping chaparral in the Santa Monica Mountains using multiple endmember spectral mixture models. Remote Sens. Environ. 65, 267–279 (1998).
Article  Google Scholar 

59.
Cavanaugh, K. C., Siegel, D. A., Reed, D. C. & Dennison, P. E. Environmental controls of giant-kelp biomass in the Santa Barbara Channel, California. Mar. Ecol. Prog. Ser. 429, 1–17 (2011).
Article  Google Scholar 

60.
Hobday, A. J. et al. A hierarchical approach to defining marine heatwaves. Prog. Oceanogr. 141, 227–238 (2016).
Article  Google Scholar 

61.
García-Reyes, M., Largier, J. L. & Sydeman, W. J. Progress in Oceanography Synoptic-scale upwelling indices and predictions of phyto- and zooplankton populations. Prog. Oceanogr. 120, 177–188 (2014).
Article  Google Scholar 

62.
McHugh, T., Abbott, D. & Freiwald, J. Phase shift from kelp forest to urchin barren along California’s North Coast. (Western Society of Naturalists, 2018).

63.
Rogers-Bennett, L., Kashiwada, J. V., Taniguchi, I. K., Kawana, S. K. & Catton, C. A. Using density-based fishery management strategies to respond to mass mortality events. J. Shellfish Res. 38, 1–11 (2019).
Article  Google Scholar 

64.
Carrascal, L. M. & Galva, I. Partial least squares regression as an alternative to current regression methods used in ecology. Oikos 118, 681–690 (2009).
Article  Google Scholar 

65.
Vadas, R. L. Ecological implications of culture studies on nereocystis luetkeana. J. Phycol. 8, 196–203 (1972).
Google Scholar 

66.
Finger, D. J. I., McPherson, M. L., Houskeeper, H. F. & Kudela, R. M. Mapping bull kelp canopy in northern California using Landsat to enable long-term monitoring. Remote Sens. Environ. 254, 112243 (2021).
Article  Google Scholar  More

1.
Pringle, R. M. A mountain of ecological interactions. Nature 568, 38–39 (2019).
CAS  Article  ADS  Google Scholar 
2.
Sayre, R. et al. A new high-resolution map of world mountains and an online tool for visualizing and comparing characterizations of global mountain distributions. Mt Res Dev 38, 240–249. https://doi.org/10.1659/mrd-journal-d-17-00107.1 (2018).
Article  Google Scholar 

3.
Peters, M. K. et al. Climate-land-use interactions shape tropical mountain biodiversity and ecosystem functions. Nature 568, 88–92. https://doi.org/10.1038/s41586-019-1048-z (2019).
CAS  Article  PubMed  ADS  Google Scholar 

4.
Bian, J., Li, A., Lei, G., Zhang, Z. & Nan, X. Global high-resolution mountain green cover index mapping based on Landsat images and Google Earth Engine. ISPRS J Photogramm Remote Sens 162, 63–76. https://doi.org/10.1016/j.isprsjprs.2020.02.011 (2020).
Article  ADS  Google Scholar 

5.
Antonelli, A. et al. Geological and climatic influences on mountain biodiversity. Nat Geo 11, 718–725. https://doi.org/10.1038/s41561-018-0236-z (2018).
CAS  Article  ADS  Google Scholar 

6.
Price, M. F., Arnesen, T., Gløersen, E. & Metzger, M. J. Mapping mountain areas: learning from Global, European and Norwegian perspectives. J Mt Sci 16, 1–15. https://doi.org/10.1007/s11629-018-4916-3 (2019).
Article  Google Scholar 

7.
Körner, C., Paulsen, J. & Spehn, E. M. A definition of mountains and their bioclimatic belts for global comparisons of biodiversity data. Alp Bot 121, 73–78. https://doi.org/10.1007/s00035-011-0094-4 (2011).
Article  Google Scholar 

8.
Körner, C. et al. A global inventory of mountains for bio-geographical applications. Alp Bot 127, 1–15. https://doi.org/10.1007/s00035-016-0182-6 (2017).
Article  Google Scholar 

9.
Karagulle, D. et al. Modeling global Hammond landform regions from 250-m elevation data. Trans GIS 21, 1040–1060. https://doi.org/10.1111/tgis.12265 (2017).
Article  Google Scholar 

10.
Payne, D., Spehn, E. M., Snethlage, M. & Fischer, M. Opportunities for research on mountain biodiversity under global change. Curr Opin Environ Sustain 29, 40–47. https://doi.org/10.1016/j.cosust.2017.11.001 (2017).
Article  Google Scholar 

11.
Silveira, F. A. O. et al. Tropical mountains as natural laboratories to study global changes: A long-term ecological research project in a megadiverse biodiversity hotspot. Perspect Plant Ecol Evol Syst 38, 64–73. https://doi.org/10.1016/j.ppees.2019.04.001 (2019).
Article  Google Scholar 

12.
Elsen, P. R., Monahan, W. B. & Merenlender, A. M. Topography and human pressure in mountain ranges alter expected species responses to climate change. Nat Commun 11, 1974. https://doi.org/10.1038/s41467-020-15881-x (2020).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

13.
Cheng, G. et al. Rebirth after death: forest succession dynamics in response to climate change on Gongga Mountain, Southwest China. J Mt Sci 15, 1671–1681. https://doi.org/10.1007/s11629-017-4435-7 (2018).
Article  Google Scholar 

14.
Trant, A., Higgs, E. & Starzomski, B. M. A century of high elevation ecosystem change in the Canadian Rocky Mountains. Sci Rep 10, 9698. https://doi.org/10.1038/s41598-020-66277-2 (2020).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

15.
Strinella, E., Scridel, D., Brambilla, M., Schano, C. & Korner-Nievergelt, F. Potential sex-dependent effects of weather on apparent survival of a high-elevation specialist. Sci Rep 10, 8386. https://doi.org/10.1038/s41598-020-65017-w (2020).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

16.
Durán-Romero, C., Medina-Sánchez, J. M. & Carrillo, P. Uncoupled phytoplankton-bacterioplankton relationship by multiple drivers interacting at different temporal scales in a high-mountain Mediterranean lake. Sci Rep 10, 350. https://doi.org/10.1038/s41598-019-57269-y (2020).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

17.
Mair, D. et al. Fast long-term denudation rate of steep alpine headwalls inferred from cosmogenic 36Cl depth profiles. Sci Rep 9, 11023. https://doi.org/10.1038/s41598-019-46969-0 (2019).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

18.
Liu, H., Chen, Y., Ye, Z., Li, Y. & Zhang, Q. Recent lake area changes in Central Asia. Sci Rep 9, 16277. https://doi.org/10.1038/s41598-019-52396-y (2019).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

19.
Spandre, P. et al. Climate controls on snow reliability in French Alps ski resorts. Sci Rep 9, 8043. https://doi.org/10.1038/s41598-019-44068-8 (2019).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

20.
Tichavský, R., Ballesteros-Cánovas, J. A., Šilhán, K., Tolasz, R. & Stoffel, M. Dry spells and extreme precipitation are the main trigger of landslides in Central Europe. Sci Rep 9, 14560. https://doi.org/10.1038/s41598-019-51148-2 (2019).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

21.
Stoffel, M., Ballesteros Cánovas, J. A., Luckman, B. H., Casteller, A. & Villalba, R. Tree-ring correlations suggest links between moderate earthquakes and distant rockfalls in the Patagonian Cordillera. Sci Rep 9, 12112. https://doi.org/10.1038/s41598-019-48530-5 (2019).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

22.
de Haas, T., Nijland, W., de Jong, S. M. & McArdell, B. W. How memory effects, check dams, and channel geometry control erosion and deposition by debris flows. Sci Rep 10, 14024. https://doi.org/10.1038/s41598-020-71016-8 (2020).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

23.
Yousefi, S. et al. A machine learning framework for multi-hazards modeling and mapping in a mountainous area. Sci Rep 10, 12144. https://doi.org/10.1038/s41598-020-69233-2 (2020).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar  More

Identification of the risk variables and their correlations with the COVID-19 damages
We have investigated a series of factors contributing to the risk of an epidemic diffusion and its impact on the population. Among many possible, we selected the following variables: mobility index, housing concentration, healthcare density, air pollution, average winter temperature and age of population. In paragraph 1 of Methods section we motivate our choice on such variables (mainly based on epidemics literature and features of the COVID-19 outbreak), show the related data (see Table 1) and explain the adopted normalization.
The first step is, of course, to estimate to what extent the chosen normalized variables individually correlate with the main impact indicators of the COVID-19 epidemic, i.e., total cases and total deaths detected in each Italian region, cumulated up to July 14, 20204, when the first epidemic wave seemed to have finished, and the intensive care occupancy recorded on April 2, 20204, when the epidemic peak was reached. In the first two rows of Fig. 2, from panel (a) to panel (f), the spatial distributions of the six risk indicators, multiplied by the population of each region, are reported as chromatic maps and thus can be visually compared with the analogous maps of the three impact indicators, panels (g), (h) and (i) in the third row. As detailed in Table 2, in paragraph 2 of Methods section, pairwise correlations between risk indicators are, with a few exceptions, quite weak; furthermore, in Table 3, results of the linear least squares fit of each individual risk indicator to damages are reported. We found correlation coefficients ranging from 0.71 to 0.96, always higher than those observed as a function of the population, which can be considered the null model; however, the relative quadratic errors stay quite high (from 0.26 to 0.62). This suggests that some opportune combination of risk indicators could better capture the risk associated to each region. In the next paragraph, we propose a risk assessment framework aimed to this.
Figure 2

The geographical distribution of the six risk factors (a–f) can be compared with the COVID-19 total cases (g), the total deaths (h) and the intensive care occupancy (i). Cases and deaths have been cumulated up to July 14, 2020, i.e. at the end of the first epidemic wave; the intensive care data have been recorded on April 2, 2020, i.e. just before the epidemic peak. The risk indicators have been multiplied for the population of each region and normalized between 0 and 1 (the color scale for temperature has been reversed, i.e. dark colors mean low temperatures, see Methods). A concentration of dark colors in the northern regions is roughly visible for almost all the indicators and the correlations between the single factors and the damages range from 0.70 to 0.95. Maps were realized with QGIS 3.10 (https://qgis.org/en/site/). (l) Crichton’s Risk Triangle. (m) Risk Index assessment framework: risk indicators (factors) are reported in red, risk components in black.

Full size image

Definition of a risk assessment framework and calibration with COVID-19 data
Conventional risk assessment theory relies on “Crichton’s Risk Triangle”24,25, shown in panel (l) of Fig. 2. In this framework, risk is evaluated as a function of three components: Hazard, Exposure and Vulnerability. Hazard is the potential for an event to cause harm (e.g., earthquake, flooding, epidemics); Exposure measures the amount of assets exposed to harm (e.g., buildings, infrastructures, population); Vulnerability is the harm proneness of assets if exposed to hazard events (e.g., building characteristics, drainage systems, age of population). The risk is present only when all of the three components co-exist in the same place. Used for the first time in the insurance industry24, this approach has been extended to assess spatially distributed risks in many fields of disaster management, such as those related to climate change impact27,28,29,30,31 and earthquakes32.
In the present paper, we consider Hazard as the degree of diffusion of the virus over the population of an Italian region (influenced by a set of factors, related to spatial and socio-economic characteristics of the region itself); Exposure is the amount of people who might potentially be infected by the virus as a consequence of the Hazard (it should coincide with the size of the population of the region); Vulnerability is the propensity of an infected person to become sick or die (in general, it is strongly related to the age and pre-existing health conditions prior to infection). The combination of Vulnerability and Exposure provides a measure of the absolute damage (i.e., the number of ill people due to pathologies related to the virus in the region), which we called Consequences.
In paragraph 3 of Methods section we propose two models that differ in the way the risk indicators are aggregated into the three components of the Crichton’s risk triangle. In particular, we consider the E_HV model, where the effect of Hazard and Vulnerability are combined in a single affine function of the six indicators, and the E_H_V model, where Hazard and Vulnerability are considered as affine functions of, respectively, mobility index, housing concentration and healthcare density, on one hand, and air pollution, average winter temperature and age of population on the other hand (see Fig. 2 (m) for a summary). In both models the Exposure is represented by the population of each region. Furthermore, two versions of each model have been considered: an optimized one, where the weights of the risk indicators are obtained through a least-square fitting versus real COVID-19 data, and an a-priori one, where all the weights are assumed to be equal.
As shown in Tables 4 and 5 of Methods section, models based on data fitting perform better, both in terms of relative mean quadratic error and correlation coefficient, as expected. In particular, the E_H_V model fits the best. Furthermore, in agreement with the strong correlation of the variables with the targets, most coefficients are positive. Indeed, all coefficients obtained by fitting the number of cases and the intensive care occupancy are positive, and only one negative coefficient appears in each model, when fitting the number of deceased. However, the numerical value of the coefficients strongly depends on both models and targets, making these models not very robust. On the other hand, the a-priori models are independent of the targets, depending only on the choice of the variables we decided to include in the risk evaluation.
Among the two considered a-priori models, where all coefficients assume the same value, we observe that the E_H_V model produces a smaller error with respect to real COVID-19 data and better correlation coefficients than the E_HV model, thus justifying the multiplicative approach which define the risk intensity in terms of the product between Hazard and Vulnerability (we used data at April 2, 2020 for this preliminary analysis but similar results would be obtained using data at July 14, 2020). Moreover, the aggregation of risk indicators in the three components of the E_H_V model follows better our motivations to choose those indicators (as explained in Methods, paragraph 1).
Validation of the a-priori E_H_V model on COVID-19 data
Once we established the robustness of the a-priori E_H_V model, let us now build the corresponding regional risk ranking and validate the model with the regional COVID-19 data as a case study. In particular, following the scheme of Fig. 2 (m), by multiplying Exposure and Vulnerability for the k-th region, we first calculate the Consequences ((C_{k} = E_{k} cdot V_{k}), k = 1,…,20). Then, by multiplying Hazard and Consequences, we obtain the global risk index (R_{k}) for each region ((R_{k} = H_{k} cdot C_{k}), k = 1,…, 20). In this respect, the risk index can be interpreted as the product of what is related to the occurrence of causes of the virus diffusion in a given region ((H_{k})) and what is related to the severity of effects on people ((C_{k})).
In Fig. 3a we can appreciate the predictive capability of our model by looking at the a-priori risk ranking of the Italian regions, compared with the COVID-19 data4, in terms of total cases (cumulated), deaths (cumulated) and intensive care occupancy (daily, not cumulated), updated both at April 2, 2020 and July 14, 2020. The values of (R_{k}) have been normalized to their maximum value, so that Lombardia results to have (R_{k}) = 1. The average of (R_{k}) over all the regions is (R_{av} = 0.15) and can be considered approximately a reference level for the Italian country (even if, of course, it has only a relative value).
Figure 3

(a) A-priori normalized risk ranking of Italian regions, emerging from our analysis of risk indicators, compared with the corresponding total cases, deaths and intensive care occupancy updated, respectively, at April 2, 2020 (just before the epidemic peak) and at July 14, 2020 (at the end of the first wave). Regions are organized in four risk groups, corresponding to different colors: very high, high, medium and low risk. The agreement with the observed effects Data referring to overestimations or underestimations of risk are also colored in green and red, respectively. (b–d) Comparison between the spatial distribution of COVID-19 total cases at July 14, 2020 (b), the most struck regions (in terms of severe cases and deaths) from 2019–2020 seasonal flu (d) according to the ISS data19 and our a-priori risk map (c). The geographical correlation with the risk map is evident for both kind of epidemic flus. Maps were realized with QGIS 3.10 (https://qgis.org/en/site/).

Full size image

As already explained, due to the intrinsic limitations of the official COVID-19 data, it is convenient to make the comparison at the aggregate level of groups of regions, without expecting to predict the exact rank within each group. Let us therefore arrange the 20 regions in four risk groups, each one characterized by a different color and ordered according to decreasing values of the risk index: very high risk ((0.4 < R_{k} le 1), in red), high risk ((0.2 < R_{k} le 0.4), in brown), medium risk ((0.03 < R_{k} le 0.2), in beige) and low risk ((R_{k} le 0.03), in pink). With this choice, our model is clearly able to correctly identify the four northern regions where the epidemic effects have been far more evident, in terms of cases, deaths and intensive care occupancy: the first in the ranking, i.e. Lombardia (whose risk score is about three times the second classified) and the group of the three regions immediately after it, Veneto, Piemonte and Emilia Romagna (even if not in the exact order of damage). A quite good agreement can be observed also for the other two groups: only for Sardegna the effects on both total cases and deaths seem to have been slightly overestimated (its insularity might play a role), while for other two regions, Umbria and Valle d’Aosta, some impact indicators have been slightly underestimated. Notice that the proposed risk classification seems quite robust, since it holds both near to the peak of April and at the end of the first wave, in July, when the intensive care occupancy of the majority of the regions was zero. In Table 6 reported in Methods, a further analysis of the robustness of this classification has been performed by eliminating, one by one, single indicators from the risk index definition: results show that the position of some regions slightly changes inside each group, but the composition of the four risk groups remains for the mostly unchanged with just few exceptions worsening the agreement with the impact indicators shown in Fig. 3a. This confirms the advantage of including all indicators in the risk index. The clear separation between northern regions from central and southern ones is also confirmed in the bottom part of Fig. 3, where the a-priori risk color map, in panel (c), is compared with the map of COVID-19 total cases in July, panel (b), and the map of the serious cases and deaths of the seasonal flu 2019/20 in Italy, panel (d) (ISS data19). The agreement is clearly visible. In Fig. 4 we show the correlations between the a-priori risk index and the three main impact indicators related to the outbreak, i.e. the total number of cases (a) and the total number of deaths (b), cumulated up to July 14, 2020, and the intensive care occupancy (c), registered at April 2, 2020. For each plot, a linear regression has been performed, with Pearson correlation coefficients always taking values greater or equal to 0.97, indicating a strong positive correlation. On the right of each plot we report the corresponding percentages of damage observed in the three Italian macro-regions—North, Center and South, see the geographic map (d). Also in this case the correlation is evident, if compared with the percentage of cumulated a-priori risk associated to the same macro-regions (e). Figure 4 The three main impact indicators for COVID-19—the total number of cases (a) and the total number of deaths (b) cumulated up to July 14, 20204, and the intensive care occupancy (c) at April 2, 20204—are reported as function of the a-priori risk index for all the Italian regions. The size of the points is proportional to the risk index score. A linear regression has been performed for each plot. The Pearson correlation coefficients are very good, always greater or equal than 0.97. The corresponding percentages of damages, aggregated for the three Italian macro-regions (North, Center and South (d)) are also reported to the right and can be compared with the percentages of cumulated a-priori risk (e). It is clear that our a-priori risk index is able to explain the anomalous damage discrepancies between these different parts of Italy. Maps were realized with QGIS 3.10 (https://qgis.org/en/site/). Full size image Another interesting way to visualize these correlations is to represent the a-priori risk index through its two main aggregated components, Hazard and Consequences, and plotting each region as a point of coordinates ((H_{i} ,C_{i} )) in the plane (left{ {H times C} right}). This Risk Diagram is reported in Fig. 5a, where the points have been also characterized by the same color of the corresponding risk group of Fig. 3. It is evident that the iso-risk line described by the equation C = Rav/H (being Rav = 0.15 the average regional risk value) is correctly able to separate the four more damaged and highly risky, northern regions (plus Lazio) from all the others. The value of the risk index is reported in parentheses next to each region name. As shown in Fig. 5b, where the ranking of the Italian regions has been disaggregated for both Hazard and Consequences, it is interesting to notice that some regions (such as Friuli, Trentino or Valle d’Aosta) exhibit high values of Hazard and quite low values of Consequences, while for other regions (such as Campania or Piemonte) the opposite is true. See also the colored geographic maps in Fig. 5c,d for a visual comparison. This confirms that it is necessary to aggregate such two main components in a single global index to have a more reliable indication of the regional a-priori risk. Figure 5 (a) Risk Diagram. Each region is represented as a point in the plane (left{ {H times C} right}) while the color is proportional to the corresponding risk group updated at July 14, 2020 (see Fig. 3a). The most damaged regions lie with a good approximation above the C = Rav/H hyperbole (i.e. the iso-risk line related to the average regional risk index), while the less damaged ones lie below this line. The a-priori risk index score is also reported for each region. (b) The rankings of Italian regions according to either Hazard (on the left) or Consequences (on the right). The corresponding colored geographic maps are also shown in panels (c) and (d) for comparison. Maps were realized with QGIS 3.10 (https://qgis.org/en/site/). Full size image Let us close this paragraph by showing, in Fig. 6, three sequences of the geographic distribution of the total cases (a), total number of deaths (b) and current intensive care occupancy (c) as a function of time, from March 9 to July 14, 2020. These sequences are compared with the geographic map of the a-priori risk level (the bordered image on the right in each sequence), the latter being independent of time. In all the plots, damages seem to spread over the regions with a variable intensity (expressed by the color scale) quite correctly predicted by our a-priori risk analysis. The intensive care occupancy map compared with the risk map is dated April 2, since the occupancy on July 14 is zero almost everywhere (with the exception of Lombardia and a few other regions). Figure 6 The geographic distributions of damage in the various Italian regions—cumulated total cases (a), cumulated total deaths (b) and daily intensive care occupancy (c)—are reported as function of time, from March 9, 2020 to July 14, 2020 and compared with the geographic distribution of the a-priori risk. Obviously, the intensive care occupancy to compare with the risk map is that of April, since in July, at the end the epidemic wave, this variable is zero everywhere except for a few regions (among which only Lombardia has a score slightly higher than 25). Maps were realized with QGIS 3.10 (https://qgis.org/en/site/). Full size image In the next paragraph, the methodology proposed in this paper, and in particular this representation in terms of risk diagram, will be used to build a policy model aimed at mitigating damages in case of an epidemic outbreak similar to the COVID-19 one. A proposal for a policy protocol to reduce the epidemic risk We have seen how the risk can be thought as composed in two components, one related to the causes of the infection diffusion and the other to the consequences. In this paragraph we will interpret the consequences in terms of protection and required support to people with the goal of improving the social result and/or reducing the economic cost. It is evident that enhancing the capability of the healthcare system appears to be the most important action: basically, the insufficient carrying capacity creates the emergency. Beyond specific factors explained above, the epidemic crisis in Lombardia essentially showed a breakdown of its healthcare system, caused by high demand rate for hospital admissions, long permanence times in intensive care, insufficient health assistance (diagnosis equipment, staff, spaces, etc.). Previously illustrated data provide a positive analysis of an epidemic disease (i.e., how things are, in a given state of the world). The normative approach here described presents a viable framework to assess possible policy protocols. Several variables affecting the diffusion of an infection can be looked at as suitable policy instruments to manage both the spreading process and the stress level to the healthcare system of a given district (such as a country, a region, an urban area, etc.). Following the evidence suggested by data, we propose a theoretical model (whose details are presented in the Methods section, paragraph 4) based on two independent variables influencing the level of risk, namely the infection ratio, i.e., the proportion of infected individuals over the total population, and the number of per capita hospital beds, as a measure of the impact of consequences caused by the spreading of the disease. We adopt an approach based on a standard model of economic policy, in which a series of instruments explicitly affecting the infection ratio and the per capita hospital beds endowment can be used to approach the target, i.e., the minimization of the risk level. A similar rationale, covering other topics, can be found in Samuelson and Solow33 (1960) and builds upon a widely consolidated literature which dates back in time34,35,36,37,38,39 (among many others). Despite the analysis concerns a collective problem, the model here proposed describes elements of a possible decision process followed by an individual policy-maker, thus remaining microeconomic in nature. Panel (a) in Fig. 7 shows the risk function, while the right panel provides an illustration of the family of its convex contours, for a finite set of risk levels (limited for graphic convenience): Figure 7 (a,b) The Risk function and its convex contours: an example for (R = x^{0.5} b^{0.5}). (c,d) The carrying capacity function and effects of policy interventions on the supply-side. (e,f) Comparative statics of equilibrium and disequilibrium. (g,h) Two examples of model implementation, see the main text. Full size image Panel (b) in Fig. 7 replicates the meaning of Fig. 5a by translating the consequences indicated by data as the required per capita hospital beds, while explaining that the position of each iso-risk curve corresponds to the different actual composition of the scenario at hand. We assume a unique care strategy based on the structural carrying capacity of the healthcare system, defined as the available number of per capita hospital beds. Such a carrying capacity derives from the health expenditure (G_{H}), which is set to a level considered sufficient. Such a choice is based on political decisions and is reasonably inferred from past experience, structural elements of population, such as age and territorial density, etc. A part of the deliberated budget is dedicated to set up intensive care beds, as an advanced assistance service provision. During an emergency, possibly deriving from an epidemic spreading, the number of beds can suddenly reveal insufficient. In other words, it is possible that the amount of hospital beds required at a certain point is greater than the current availability. In the model, we assume the number of hospital beds, H, and the proportion of intensive care beds, (alpha), as exogenously determined by the policy-maker who fixes (G_{H}). The actual carrying capacity is shown as a function of the infection ratio, x, computed as the infected population over the total, as shown in panel (c) of Fig. 7, and detailed in paragraph 4 of Methods. Changes in the proportion of per capita intensive care hospital beds over the total, cause instead, a variation in the slope of the line (which becomes steeper for reduction in the proportion of intensive care beds). Finally, changes in the overall expenditure shift the line with the same slope (above for increments of the expenditure). In particular, it is worth to notice that the political choice of the ratio (alpha = HH/H) may imply that the overall capacity to assist the entire population is not guaranteed (i.e. the intercept on the (x) axis might be less than (1)). A direct comparison of elements contained in panels (a-b) and (c-d) of Fig. 7 provides a quick inspection of the policy problem, focused to control the epidemic spreading. The constraint should be considered as a dynamic law, but since the speed of adjustment is reasonably low, we will proceed by means of a comparative statics perspective, in which a comparison of different strategies can be presented, by starting from different, static, scenarios. Further, by definition, an emergency challenges the usual policy settings, since the speed of damages is greater than that of policy tools. In panel (e) of Fig. 7 a hypothetic country has a given carrying capacity to sustain the risk level represented by the iso-risk curve. Without an immediate availability of funds to increase the carrying capacity, the main policy target could easily be described as the transposition of the iso-risk curve to the bottom-left: the closer the curve to the origin, the higher the satisfaction for the community. Secondly, the meaning of the relationship between the curve and the line is that until the curve touches the line, the policy maker has a sort of measure of how much the problem is out of control, given by the distance between the curve and the constraint. Third, policies may try to transpose the curve to lower levels or, equivalently, the constraint upwards (with or without modification of the slope). A minimal result is reached if both are at least tangent, as depicted in panel (f) of Fig. 7. Whenever such a tangency condition has been reached, the highest infection rate that the given health care system can sustain has been found. Further policy actions are possible to approach a lower iso-risk curve or to save resources and/or re-allocate them differently. A policy can be considered satisfactory when any of points belonging to the arc TT’ is reached, e.g. the point L. Alternative policies are neither equivalent, nor requiring the same actions, and the policy-maker has to choose actions with reference to the actual data collected by its own Country. Points F and G, although carrying the same risk level as E, still represent out-of-control positions. Different regions of the plot have a different signaling power: at point F, the infection rate is low and, thus, very difficult to be further reduced. In such a case, for example, it would be advisable to suggest health protocols which improve people safety. On the contrary, at point G, the infection rate is so high that a limit on social interaction easily appears to be much more urgent than medical protocols. The right mix between a demand-side and a supply-side policy to adopt is a decision of political nature. A distinction can be made by saying that demand-side policies are devoted to reduce the number of newly infected people (by means of restrictions to movements, quarantine regulations, rules of conduct, etc.) and their effects are able to lower the iso-risk curves; supply-side policies are, instead, aimed at incrementing the carrying capacity of the system (by means of expenditure for the healthcare system, increments of dedicated personnel and intensive care beds, in-house medical protocols) and their effects can shift the constraint representing the carrying capacity of the system. Politics has, then, to decide when the risk is low enough or the constraint is sufficiently high. Specific calibration of the model will allow, in a forthcoming research, a detailed analysis of policy implications, by considering actual conditions and risk factors of specific districts, thus providing the policy-maker with a toolbox for normative directions. For instance, the model can be read to analyze differences in proposed actions in Lombardia and Veneto, and in other regions or countries. More

Photographs of crabs and backgrounds
We sampled crabs and backgrounds to obtain images for the game. The population used was located in Falmouth (50.141888, −5.063811) on the south coast of the UK, comprising a stretch of shoreline encompassing neighbouring Castle and Gyllyngvase beaches. The crab habitats at the site comprise rock pools with rocky crevices with stony or gravel substrates in the pools and, lower down on the shore, increasing abundance of seaweed21. Together these create visually variable textures and heterogeneity in crab habitat types.
Photographs of natural backgrounds (rock pools) were taken by Samsung NX1000 digital camera converted to full spectrum and attached with a Nikon EL 80 mm lens. Background sampling was conducted along three ~100 m long transects placed parallel to the shoreline across different tide-zones (i.e. low, middle, high) spaced evenly down the beach (following21). Each of the backgrounds photographed were at least 5 m apart from each other (i.e. transect was subdivided approximately into 5-m-intervals) ensuring the variability in background types across transect. These sampling quadrats were photographed during low-tide to avoid specular light reflecting back from the water. To obtain images that capture naturalistic colour variation, the images were taken in RAW format with manual white balance and a fixed aperture setting. For human visible photos as used here, we placed a UV and infra-red (IR) blocking filter in front of the lens, which transmits wavelengths only between 400–680 nm (Baader UV/IR Cut Filter). We have previously characterised the spectral sensitivity of our cameras39. For calibration purposes, each photograph included a grey reflectance standard, which reflects light equally at 7 and 93% between 300 and 750 nm.
Quadrats were searched for shore crabs for a period of ~5 min. We searched for crabs by raking gravel by hand, moving small boulders aside, turning seaweed over and checking crevices to ensure any crabs were unlikely to be missed. After crabs were found we transported them to laboratory facilities at the University of Exeter Penryn campus for standardised photography. During the transportation all crabs were kept on standard average grey buckets. Photographs of crabs were taken with the same camera set up as above. In the laboratory a bulb simulating D65 illuminant (Iwasaki eyeColor bulb) was used while crabs were photographed against grey standard background. We included grey standards and scale bars in the photographs. Images were then calibrated and converted to normalised reflectance images (relative to the grey standard)39,40.
Crab images were scaled into the same pixel/mm aspect ratio to show crabs against the background images in natural size with respect to the background scale. Following past work25, crab outlines were cut out from the image by custom software was designed (called ‘autocrab’) to automate the process of background subtraction. This software allowed us to step through hundreds of images, automatically loading, thresholding and flood filling background areas, saving them with an appropriate transparency channel in the correct format and resolution needed for the game. This created usable crab images for 80% of the photographs easily, with some additional cleaning up required for the rest using GIMP2 image manipulation software (https://zenodo.org/record/1101057; DOI for the source code: https://doi.org/10.5281/zenodo.1099634). The crab images were PNGs (portable network graphic) with a variable alpha level to ensure there were no jagged edges visible.
Selection of crabs
We aimed to ensure that we had an ecologically relevant range of crab phenotypes used in the game. We also sought to test how different types or ‘morphs’ of crab would affect search image formation and detection. Therefore, we used a procedure to categorise crabs into one of six categories prior the experiment. Note that, statistically crab variation may be more continuous rather than falling into true morphs, but there are a number of common crab patterns and features that frequently arise in the wild20, potentially reflecting ‘modules’ of development and pattern expression. We emphasise that our aim here was not to test specifically whether shore crabs occur in discrete morphs, but rather to capture some of the variation and common features that exist in this species in order to explore the effects of different pattern types on search image formation and whether effects differ among common categories of appearance.
Game design
The design of the experiment generally followed the approach of previous citizen science camouflage games24. Ethical approval was granted by Exeter University (ID: 2015/736). Subjects were recruited via social media and word of mouth. On loading the webpage, subjects were taken to a start screen and informed that the game was an experiment and that by playing they consented to their data being used. They were free to leave the game at any time and no personal or identifying data were collected. Subjects also asked if they had played the game before.
The game was programmed in HTML5 (including JavaScript, CSS and PHP), and was available to play on all standard internet browsers. Upon loading the game each participant was shown a series of photographs of 24 natural rock pool backgrounds (randomly sampled from 105 natural background images) with a single crab (randomly sampled from 155 natural crab images) in each image (Fig. 1). Participants were asked to detect the crab (by clicking on it) as quickly as possible, which would progress them to the next slide. If the crab was not found within 15 s the crab was highlighted with a circle for 1 s, and then the participant progressed to the next slide. During the experiment, the probability of being shown the same individual crab phenotype in the next slide was always 80% (although the crab’s position and rotation, and the background image were all randomised), meaning that subjects were likely to have runs of the same individual crab in succession, often up to 10 encounters (the median run length for each crab being ~5 encounters). This approach mimicked a situation where there is no intraspecific variation in pattern, and allowed us to test which aspects of crab/morph appearance affected search image formation and switching.
Analysis of crab appearance and camouflage
Following our previous work testing how different types of camouflage metric predict detection26, we analysed a large number of metrics linked to camouflage efficacy, these include edge disruption, colour, luminance (lightness), and pattern metrics. The metrics included crab-only appearance measures (such as the crab’s intrinsic colour, brightness, and dominant marking size), and also comparative metrics where each crab is compared to its local surroundings (within a radius of one body-length, where body length is described as the diameter of a circle which best fits the crab’s outline), and also the crab compared to the entire background image. In total there were 45 metrics, all described in Supplementary Data 1. All image analysis was performed using ImageJ v1.5041, code available on request.
Images were converted from sRGB to CIELAB colour space before measuring them given that humans were the participants used in this study. Each crab was measured by recreating its exact position and rotation on each background for image analysis.
Luminance distribution difference was measured from the CIE L channel in 100 bins following the methods described in Troscianko et al.26, effectively the sum of absolute differences between the crab’s luminance histogram and the background or surrounding’s luminance histogram. The highly variable nature of the crab’s colour and background colours mean that calculating a mean colour for the background or crab may not be appropriate because it creates intermediate colours which do not represent the scene as a whole. Therefore, a colour equivalent of the luminance distribution difference method was also developed, where pixel CIE A and B values were plotted in a two-dimensional histogram to create a proportional frequency “map”. Each axis had 200 bins ranging from −100 to 100, meaning the bins are smaller than the human colour discrimination threshold in CIE LAB space. The absolute differences in the crab’s colour map and its background or surround colour maps were used as a non-parametric method for describing background colour matching. Edge disruption was also measured following the GabRat approach described in Troscianko et al. (2017), however in addition to measuring the CIE L image, the chromatic opponent channel images (CIE A and B images) were also measured (i.e. as a measure of chromatic edge disruption). Pattern energy difference was measured by creating a series of bandpass images, filtering each crab and surround into different spatial scales, then measuring the degree of “energy” standard deviation in pixel values) at each spatial scale to create an energy spectrum. Pattern energy difference calculates the absolute sum of energy differences at each spatial scale between the crab and its background following Troscianko et al.26.
Statistics and reproducibility
Survival models were used to determine how crab capture times were affected by experimental treatments and camouflage variables. Survival models offer the ability to count crabs reaching “timeout” (where participants still could not find the crab after 15 s) as surviving up to this point (termed censored in survival models). Mixed effects survival models (coxme version 2.2–1027) were used to reflect the fact that within-session data are not independent. All statistical analyses were performed in R (version 3.4.4), with the raw data and R script available as supplementary material (“Supplementary Data 2”, and “Supplementary Data 3” respectively). We used four different models to test each of our key predictions: (i) models ranking each of the camouflage metrics in order to find the best predictor of human performance, within each camouflage strategy the best predictor was selected and used in the subsequent tests; (ii) models testing the rate of improvement in capture time for each phenotype; (iii) models comparing the capture time and appearance of each crab relative to those of the previously encountered crab; (iv) models comparing the capture time of each crab given its morph, and the morph of the previous crab (i.e. interaction between individual phenotype and overall morph). We describe each in turn here:
First, based on our metrics of camouflage, we worked out the best predictor of human performance within each of these metrics. An example of the survival model is:
coxme(Surv(cTime, hit) ~ screenScale + playedBefore + poly(crab_circular_fit_centre_x,2) + poly(crab_circular_fit_centre_y,2) + L_GabRat_sig2.0 + crab_area + (1|sessionID), data).
This model takes into account the screen resolution, whether subjects have played before, the slide number (learning within session), the screen coordinates of the crabs (crabs in the corners of the screen take longer to find), the camouflage metric (GabRat luminance edge disruption in this example), the size of the crab (bigger crabs are easier to find), and session ID as a random factor. From these models we could calculate the metrics that were most effective in predicting detection times26, and narrowed the metrics down to the best predictors of luminance, colour, pattern and edge disruption.
Second, we tested how the number of previous encounters with the current crab phenotype affected capture times. This is testing for speed-of-improvement within each phenotype, and how different types of camouflage (determined above) affect this. An example survival model is:
coxme(Surv(cTime, hit) ~ screenScale + playedBefore + slide + poly(crab_circular_fit_centre_x,2) + poly(crab_circular_fit_centre_y,2) + L_GabRat_sig2.0 * encounters + crab_area + (1|sessionID), data). Where ‘encounters’ codes for the number of previous encounters with the current phenotype.
Third, we tested capture time differences when switching between crabs, comparing the camouflage of the previous crab with the current one (note the previously encountered crab was sometimes the same phenotype, and sometimes would switch to a new one). The dependent variable (timeDiff) was log(current crab capture time) – log(previous crab capture time). The camouflage variables are calculated in the same manner, e.g. the current level of disruption minus the previous level of disruption. Here, an interaction with the number of prior encounters with the current crab phenotype shows how switching is affected by prior experience of this camouflage type. An example model is:
lmer(timeDiff ~ crab_area + pArea + playedBefore + slide + poly(crab_circular_fit_centre_x,2) + poly(crab_circular_fit_centre_y,2) + poly(pX,2) + poly(pY,2) + drpLDiff*novelCrab + (1|sessionID), diffData). The values pArea, pX and pY denote the size and screen location of the previous crab.
Finally, we analysed capture time differences when switching between each of the six crab morphs (rather than comparing camouflage metric differences), using the timeDiff value as above. An example model is:
lmer(timeDiff ~ crab_area + pArea + slide + poly(crab_circular_fit_centre_x,2) + poly(crab_circular_fit_centre_y,2) + poly(pX,2) + poly(pY,2) + slide + morphSwitch*novelCrab + (1|sessionID), morphData). Here ‘morphSwitch’ has two levels which describe whether a switch event was to the same, or a different morph. The random factor ‘sessionID’ explained almost zero variance in this dataset, and where this occurred the models were cross-validated with GLMs (see Supplementary Data 3).
Selection of crab phenotypes
We asked 10 naïve participants (who had no prior experience of crab phenotype discrimination) to subjectively sort images of crabs into distinct categories. People were not instructed on how many groups they should form – they were simply asked to group crabs based on their colour and patterning (i.e. phenotypic variation). This resulted in six categories (the actual numbers of the crab images representing that phenotype are given in brackets as follows): Black (22), Disruptive (15), Green (50), Mottled (28), Pale (20) and Spotted (20). Although this is subjective, we subsequently analysed the appearance of crabs from these categories and showed that ‘crab morph’ is a significant predictor of a range of appearance metrics, including colour, luminance, mean pattern energy, and dominant marking size (P  More

1.
Tsukamoto, K. Discovery of the spawning area for Japanese eel. Nature 356, 789–791 (1992).
ADS  Article  Google Scholar 
2.
Tsukamoto, K. Spawning of eels near a seamount. Nature 439, 929 (2006).
ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

3.
Chow, S. et al. Discovery of mature freshwater eels in the open ocean. Fish. Sci. 75, 257–259 (2009).
CAS  Article  Google Scholar 

4.
Kurogi, H. et al. First capture of post-spawning female of the Japanese eel Anguilla japonica at the southern West Mariana Ridge. Fish. Sci. 77, 199–205 (2011).
CAS  Article  Google Scholar 

5.
Tsukamoto, K. et al. Positive buoyancy in eel leptocephali: an adaptation for life in the ocean surface layer. Mar. Biol. 156, 835–846 (2009).
Article  Google Scholar 

6.
Cheng, P. W. & Tzeng, W. N. Timing of metamorphosis and estuarine arrival across the dispersal range of the Japanese eel Anguilla japonica. Mar. Ecol. Prog. Ser. 131, 87–96 (1996).
ADS  Article  Google Scholar 

7.
Chen, J. Z., Huang, S. L. & Han, Y. S. Impact of long-term habitat loss on the Japanese eel Anguilla japonica. Estuar. Coast. Shelf Sci. 151, 361–369 (2014).
ADS  CAS  Article  Google Scholar 

8.
Tanaka, E. Stock assessment of Japanese eels using Japanese abundance indices. Fish. Sci. 80, 1129–1144 (2014).
CAS  Article  Google Scholar 

9.
Jacoby, D. & Gollock, M. Anguilla anguilla. The IUCN red list of threatened species, version 2014.2. IUCN 2014 e.T60344A45833138. https://doi.org/10.1108/ICS-04-2017-0025 (2014).

10.
Onda, H. et al. Vertical distribution and assemblage structure of leptocephali in the North Equatorial Current region of the western Pacific. Mar. Ecol. Prog. Ser. 575, 119–136 (2017).
ADS  Article  Google Scholar 

11.
Saijo, Y., Iizuka, S. & Asaoka, O. Chlorophyll maxima in Kuroshio and adjacent area. Mar. Biol. 4, 190–196 (1969).
CAS  Article  Google Scholar 

12.
Furuya, K. Subsurface chlorophyll maximum in the tropical and subtropical western Pacific Ocean: Vertical profiles of phytoplankton biomass and its relationship with chlorophylla and particulate organic carbon. Mar. Biol. 107, 529–539 (1990).
CAS  Article  Google Scholar 

13.
Otake, T., Nogami, K. & Maruyama, K. Dissolved and particulate organic matter as possible food sources for eel leptocephali. Mar. Ecol. Prog. Ser. 92, 27–34 (1993).
ADS  Article  Google Scholar 

14.
Mochioka, N. & Iwamizu, M. Diet of anguilloid larvae: Leptocephali feed selectively on larvacean houses and fecal pellets. Mar. Biol. 125, 447–452 (1996).
Google Scholar 

15.
Miller, M. J., Otake, T. & Aoyama, J. Observations of gut contents of leptocephali in the North Equatorial current and Tomini Bay Indonesia. Coast. Mar. Sci. 35, 277–288 (2012).
Google Scholar 

16.
Tomoda, T. et al. Observations of gut contents of anguilliform leptocephali collected in the western North Pacific. Nippon Suisan Gakkaishi 84, 32–44 (2018).
Article  Google Scholar 

17.
Deibel, D., Parrish, C. C., Grønkjær, P., Munk, P. & GisselNielsen, T. Lipid class and fatty acid content of the leptocephalus larva of tropical eels. Lipids 47, 623–634 (2012).
CAS  PubMed  Article  PubMed Central  Google Scholar 

18.
Liénart, C. et al. Geographic variation in stable isotopic and fatty acid composition of anguilliform leptocephali and particulate organic matter in the South Pacific. Mar. Ecol. Prog. Ser. 544, 225–241 (2016).
ADS  Article  CAS  Google Scholar 

19.
Miller, M. J. et al. A low trophic position of Japanese eel larvae indicates feeding on marine snow. Biol. Lett. 9, 20120826 (2013).
PubMed  PubMed Central  Article  Google Scholar 

20.
Miyazaki, S. et al. Stable isotope analysis of two species of anguilliform leptocephali (Anguilla japonica and Ariosoma major) relative to their feeding depth in the North Equatorial Current region. Mar. Biol. 158, 2555–2564 (2011).
CAS  Article  Google Scholar 

21.
Chow, S. et al. Japanese eel Anguilla japonica do not assimilate nutrition during the oceanic spawning migration: evidence from stable isotope analysis. Mar. Ecol. Prog. Ser. 402, 233–238 (2010).
ADS  CAS  Article  Google Scholar 

22.
Chow, S. et al. Onboard rearing attempts for the Japanese eel leptocephali using POM-enriched water collected in the Western North Pacific. Aquat. Living Resour. 30, 1–7 (2017).
Article  CAS  Google Scholar 

23.
Miller, M. J., Hanel, R., Feunteun, E. & Tsukamoto, K. The food source of Sargasso Sea leptocephali. Mar. Biol. 167, 57 (2020).
CAS  Article  Google Scholar 

24.
Pompanon, F. et al. Who is eating what: Diet assessment using next generation sequencing. Mol. Ecol. 21, 1931–1950 (2012).
CAS  PubMed  Article  PubMed Central  Google Scholar 

25.
Wang, M. & Jeffs, A. G. Nutritional composition of potential zooplankton prey of spiny lobster larvae: a review. Rev. Aquac. 6, 270–299 (2014).
Article  Google Scholar 

26.
Ho, T. W., Hwang, J. S., Cheung, M. K., Kwan, H. S. & Wong, C. K. Dietary analysis on the shallow-water hydrothermal vent crab Xenograpsus testudinatus using Illumina sequencing. Mar. Biol. 162, 1787–1798 (2015).
CAS  Article  Google Scholar 

27.
Chow, S. et al. Molecular diet analysis of Anguilliformes leptocephalus larvae collected in the western North Pacific. PLoS ONE 14, e0225610 (2019).
CAS  PubMed  PubMed Central  Article  Google Scholar 

28.
Riemann, L. et al. Qualitative assessment of the diet of European eel larvae in the Sargasso Sea resolved by DNA barcoding. Biol. Lett. 6, 819–822 (2010).
CAS  PubMed  PubMed Central  Article  Google Scholar 

29.
Ayala, D. J. et al. Gelatinous plankton is important in the diet of European eel (Anguilla anguilla) larvae in the Sargasso Sea. Sci. Rep. 8, 6156 (2018).
ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

30.
Estrada, M. et al. Phytoplankton across tropical and subtropical regions of the Atlantic Indian and Pacific Oceans. PLoS ONE 11, e0151699 (2016).
PubMed  PubMed Central  Article  CAS  Google Scholar 

31.
Lundgreen, R. B. C. et al. Eukaryotic and cyanobacterial communities associated with marine snow particles in the oligotrophic Sargasso Sea. Sci. Rep. 9, 1–12 (2019).
CAS  Article  Google Scholar 

32.
Ayala, D., Riemann, L. & Munk, P. Species composition and diversity of fish larvae in the Subtropical Convergence Zone of the Sargasso Sea from morphology and DNA barcoding. Fish. Oceanogr. 25, 85–104 (2016).
Article  Google Scholar 

33.
Arai, M. N. Active and passive factors affecting aggregations of hydromedusae: a review. Sci. Mar. 56, 99–108 (1992).
Google Scholar 

34.
Boero, F. et al. Gelatinous plankton: Irregularities rule the world (sometimes). Mar. Ecol. Prog. Ser. 356, 299–310 (2008).
ADS  Article  Google Scholar 

35.
Purcell, J. E. Feeding and growth of the siphonophore Muggiaea atlantica (Cunningham 1893). J. Exp. Mar. Bio. Ecol. 62, 39–54 (1982).
Article  Google Scholar 

36.
Alldredge, A. Particle aggregation dynamics. In Encyclopedia of Ocean Sciences, 2nd edn, 330–337 (Elsevier Inc., 2008). https://doi.org/10.1016/B978-012374473-9.00468-9

37.
Hosia, A. & Bamstedt, U. Seasonal abundance and vertical distribution of siphonophores in western Norwegian fjords. J. Plankton Res. 30, 951–962 (2008).
Article  Google Scholar 

38.
Lo, W. T., Yu, S. F. & Hsieh, H. Y. Effects of summer mesoscale hydrographic features on epipelagic siphonophore assemblages in the surrounding waters of Taiwan, western North Pacific Ocean. J. Oceanogr. 69, 495–509 (2013).
Article  Google Scholar 

39.
Lo, W.-T., Yu, S.-F. & Hsieh, H.-Y. Hydrographic processes driven by seasonal monsoon system affect siphonophore assemblages in tropical-subtropical waters (Western North Pacific Ocean). PLoS ONE 9, e100085 (2014).
ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

40.
Li, K. Z., Yin, J. Q., Huang, L. M. & Song, X. Y. Comparison of siphonophore distributions during the southwest and northeast monsoons on the northwest continental shelf of the South China Sea. J. Plankton Res. 34, 636–641 (2012).
Article  Google Scholar 

41.
López-López, L., Molinero, J. C., Tseng, L.-C., Chen, Q.-C. & Hwang, J.-S. Seasonal variability of the gelatinous carnivore zooplankton community in Northern Taiwan. J. Plankton Res. 35, 677–683 (2013).
Article  Google Scholar 

42.
Price, J. F. Upper ocean response to a hurricane. J. Phys. Ocean. 11, 153–175 (1981).
ADS  Article  Google Scholar 

43.
Toratani, M. Primary production enhancement by typhoon Ketsana in 2003 in western North Pacific. In Remote Sensing of Inland, Coastal, and Oceanic Waters (eds. Frouin, R. J. et al.) 7150, 715013 (SPIE, 2008).

44.
Lin, I. I. Typhoon-induced phytoplankton blooms and primary productivity increase in the western North Pacific subtropical ocean. J. Geophys. Res. Ocean. 117, C03039 (2012).
ADS  Article  Google Scholar 

45.
Ishida, H., Furusawa, K., Makino, T., Ishizaka, J. & Watanabe, Y. The effect of typhoons on phytoplankton communities and settling particle flux in the western North Pacific subtropical region. Oceanogr. Jpn. 25, 17–41 (2016).
Article  Google Scholar 

46.
Siswanto, E., Ishizaka, J., Yokouchi, K., Tanaka, K. & Tan, C. K. Estimation of interannual and interdecadal variations of typhoon-induced primary production: a case study for the outer shelf of the East China Sea. Geophys. Res. Lett. 34, L03604 (2007).
ADS  Article  Google Scholar 

47.
Chen, Y. L. L., Houng-Yung, C., Jan, S. & Tuo, S. H. Phytoplankton productivity enhancement and assemblage change in the upstream Kuroshio after typhoons. Mar. Ecol. Prog. Ser. 385, 111–126 (2009).
ADS  CAS  Article  Google Scholar 

48.
Tsuchiya, K. et al. Typhoon-induced response of phytoplankton and bacteria in temperate coastal waters. Estuar. Coast. Shelf Sci. 167, 458–465 (2015).
ADS  CAS  Article  Google Scholar 

49.
Typhoon information. Japan Meteorological Agency. https://www.data.jma.go.jp/fcd/yoho/typhoon/index.html. Accessed 10 Dec 2020.

50.
Miller, M. J. et al. Morphology and gut contents of anguillid and marine eel larvae in the Sargasso Sea. Zool. Anz. 279, 138–151 (2019).
Article  Google Scholar 

51.
Singh, P., Liu, Y., Li, L. & Wang, G. Ecological dynamics and biotechnological implications of thraustochytrids from marine habitats. Appl. Microbiol. Biotechnol. 98, 5789–5805 (2014).
CAS  PubMed  Article  Google Scholar 

52.
Tanaka, H., Kagawa, H., Ohta, H., Unuma, T. & Nomura, K. The first production of glass eel in captivity: fish reproductive physiology facilitates great progress in aquaculture. Fish Physiol. Biochem. 28, 493–497 (2003).
Article  Google Scholar 

53.
Stenly, W. et al. Ingestion by Japanese eel Anguilla japonica larvae on various minute zooplanktons. Aquac. Sci. 61, 341–347 (2013).
Google Scholar 

54.
Butts, I. A. E., Sørensen, S. R., Politis, S. N. & Tomkiewicz, J. First-feeding by European eel larvae: a step towards closing the life cycle in captivity. Aquaculture 464, 451–458 (2016).
Article  Google Scholar 

55.
Tsukamoto, K. & Miller, M. J. The mysterious feeding ecology of leptocephali: a unique strategy of consuming marine snow materials. Fish. Sci. 87, 11–29 (2020).
Article  CAS  Google Scholar 

56.
Bouilliart, M., Tomkiewicz, J., Lauesen, P., De Kegel, B. & Adriaens, D. Musculoskeletal anatomy and feeding performance of pre-feeding engyodontic larvae of the European eel (Anguilla anguilla). J. Anat. 227, 325–340 (2015).
PubMed  PubMed Central  Article  Google Scholar 

57.
Westeberg, H. A proposal regarding the source of nutrition of leptocephalus larvae. Int. Rev. Hydrobiol. Hydrogr. 75, 863–864 (1990).
Article  Google Scholar 

58.
Miller, M. Ecology of anguilliform leptocephali: remarkable transparent fish larvae of the ocean surface layer. Aqua-BioScience Monogr. https://doi.org/10.1093/gbe/evy021 (2009).
ADS  Article  Google Scholar 

59.
Strom, S., Bright, K., Fredrickson, K. & Brahamsha, B. The Synechococcus cell surface protein SwmA increases vulnerability to predation by flagellates and ciliates. Limnol. Oceanogr. 62, 784–794 (2017).
ADS  Article  Google Scholar 

60.
Benner, R. & Kaiser, K. Abundance of amino sugars and peptidoglycan in marine particulate and dissolved organic matter. Limnol. Oceanogr. 48, 118–128 (2003).
ADS  CAS  Article  Google Scholar 

61.
Seymour, J., Ahmed, T., Durham, W. & Stocker, R. Chemotactic response of marine bacteria to the extracellular products of Synechococcus and Prochlorococcus. Aquat. Microb. Ecol. 59, 161–168 (2010).
Article  Google Scholar 

62.
Biller, S. J. et al. Bacterial vesicles in marine ecosystems. Science (80-) 343, 183–186 (2014).
ADS  CAS  Article  Google Scholar 

63.
Scanlan, D. Bacterial vesicles in the ocean. Science 343, 143–144 (2014).
ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

64.
Cisternas-Novoa, C., Lee, C. & Engel, A. Transparent exopolymer particles (TEP) and Coomassie stainable particles (CSP): Differences between their origin and vertical distributions in the ocean. Mar. Chem. 175, 56–71 (2015).
CAS  Article  Google Scholar 

65.
Long, R. A. & Azam, F. Abundant protein-containing particles in the sea. Aquat. Microb. Ecol. 10, 213–221 (1996).
Article  Google Scholar 

66.
Tanoue, E., Ishii, M. & Midorikawa, T. Discrete dissolved and particulate proteins in oceanic waters. Limnol. Oceanogr. 41, 1334–1343 (1996).
ADS  CAS  Article  Google Scholar 

67.
Simon, M., Alldredge, A. L. & Azam, F. Bacterial carbon dynamics on marine snow. Mar. Ecol. Prog. Ser. 65, 205–211 (1990).
ADS  CAS  Article  Google Scholar 

68.
Godhe, A. et al. Quantification of diatom and dinoflagellate biomasses in coastal marine seawater samples by real-time PCR. Appl. Environ. Microbiol. 74, 7174–7182 (2008).
CAS  PubMed  PubMed Central  Article  Google Scholar 

69.
Zhu, F., Massana, R., Not, F., Marie, D. & Vaulot, D. Mapping of picoeucaryotes in marine ecosystems with quantitative PCR of the 18S rRNA gene. FEMS Microbiol. Ecol. 52, 79–92 (2005).
CAS  PubMed  Article  PubMed Central  Google Scholar 

70.
Gong, W. & Marchetti, A. Estimation of 18S gene copy number in marine eukaryotic plankton using a next-generation sequencing approach. Front. Mar. Sci. 6, 219 (2019).
Article  Google Scholar 

71.
Furuya, K. & Marumo, R. The structure of the phytoplankton community in the subsurface chlorophyll maxima in the western North Pacific Ocean. J. Plankton Res. 5, 393–406 (1983).
Article  Google Scholar 

72.
Kuroki, M., Okamura, A., Yamada, Y., Hayasaka, S. & Tsukamoto, K. Evaluation of optimum temperature for the early larval growth of Japanese eel in captivity. Fish. Sci. 85, 801–809 (2019).
CAS  Article  Google Scholar 

73.
Okamura, A. et al. Effects of water temperature on early development of Japanese eel Anguilla japonica. Fish. Sci. 73, 1241–1248 (2007).
CAS  Google Scholar 

74.
Kurokawa, T. et al. Influence of water temperature on morphological deformities in cultured larvae of Japanese eel, Anguilla japonica, at completion of yolk resorption. J. World Aquac. Soc. 39, 726–735 (2008).
Article  Google Scholar 

75.
Tsukamoto, K. et al. Oceanic spawning ecology of freshwater eels in the western North Pacific. Nat. Commun. 2, 179 (2011).
ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

76.
Shirai, K. et al. Temperature and depth distribution of Japanese eel eggs estimated using otolith oxygen stable isotopes. Geochim. Cosmochim. Acta 236, 373–383 (2018).
ADS  CAS  Article  Google Scholar 

77.
Ichikawa, T. Particulate organic carbon and nitrogen in the adjacent seas of the Pacific Ocean. Mar. Biol. 68, 49–60 (1982).
CAS  Article  Google Scholar 

78.
Hebel, D. V. & Karl, D. M. Seasonal, interannual and decadal variations in particulate matter concentrations and composition in the subtropical North Pacific Ocean. Deep Sea Res Part II Top. Stud. Oceanogr. 48, 1669–1695 (2001).
ADS  CAS  Article  Google Scholar 

79.
MacIntyre, S., Alldredge, A. L. & Gotschalk, C. C. Accumulation of marines now at density discontinuities in the water column. Limnol. Oceanogr. 40, 449–468 (1995).
ADS  Article  Google Scholar 

80.
Tomas, C. R. & Hasle, G. R. Identifying Marine Phytoplankton (Academic Press, New York, 1997).
Google Scholar 

81.
Suzuki, K. et al. Responses of phytoplankton and heterotrophic bacteria in the northwest subarctic Pacific to in situ iron fertilization as estimated by HPLC pigment analysis and flow cytometry. Prog. Oceanogr. 64, 167–187 (2005).
ADS  Article  Google Scholar 

82.
Nagai, S. et al. Influences of diurnal sampling bias on fixed-point monitoring of plankton biodiversity determined using a massively parallel sequencing-based technique. Gene 576, 667–675 (2016).
CAS  PubMed  Article  PubMed Central  Google Scholar 

83.
Tanabe, A. S. et al. Comparative study of the validity of three regions of the 18S-rRNA gene for massively parallel sequencing-based monitoring of the planktonic eukaryote community. Mol. Ecol. Resour. 16, 402–414 (2016).
CAS  PubMed  Article  PubMed Central  Google Scholar 

84.
Dzhembekova, N., Moncheva, S., Ivanova, P., Slabakova, N. & Nagai, S. Biodiversity of phytoplankton cyst assemblages in surface sediments of the Black Sea based on metabarcoding. Biotechnol. Biotechnol. Equip. 32, 1507–1513 (2018).
CAS  Article  Google Scholar 

85.
Dzhembekova, N., Urusizaki, S., Moncheva, S., Ivanova, P. & Nagai, S. Applicability of massively parallel sequencing on monitoring harmful algae at Varna Bay in the Black Sea. Harmful Algae 68, 40–51 (2017).
PubMed  Article  PubMed Central  Google Scholar 

86.
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
CAS  PubMed  PubMed Central  Article  Google Scholar 

87.
Schloss, P. D., Gevers, D. & Westcott, S. L. Reducing the effects of PCR amplification and sequencing artifacts on 16S rRNA-based studies. PLoS ONE 6, e27310 (2011).
ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

88.
Edgar, R. C., Haas, B. J., Clemente, J. C., Quince, C. & Knight, R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194–2200 (2011).
CAS  PubMed  PubMed Central  Article  Google Scholar 

89.
Cheung, K. L. Y., Huen, J., Houry, W. A. & Ortega, J. Comparison of the multiple oligomeric structures observed for the Rvb1 and Rvb2 proteins. Biochem. Cell Biol. 88, 77–88 (2010).
CAS  PubMed  PubMed Central  Article  Google Scholar 

90.
Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinform. 10, 421 (2009).
Article  CAS  Google Scholar 

91.
Horton, T. et al. World register of marine species (WoRMS) (2018).

92.
R Core team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing , Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/ (2017). https://doi.org/10.2788/95827.

93.
Bray, J. R. & Curtis, J. T. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27, 325–349 (1957).
Article  Google Scholar 

94.
Anderson, M. J. A new method for non-parametric multivariate analysis of variance. Aust. Ecol. 26, 32–46 (2001).
Google Scholar 

95.
Oksanen, J. et al. vegan: Community ecology package. R package version 2.5-2. CRAN R (2018). ISBN 0-387-95457-0. More

1.
FAO. United Nations Food Agricultural Organisation. High Level Expert Forum (FAO, Rome, 2009).
Google Scholar 
2.
Puig-Montserrat, X. et al. Pest control service provided by bats in Mediterranean rice paddies: Linking agroecosystems structure to ecological functions. Mamm. Biol. 80, 237–245. https://doi.org/10.1016/j.mambio.2015.03.008 (2015).
Article  Google Scholar 

3.
Cleveland, C. J. et al. Economic value of the pest control service provided by Brazilian free-tailed bats in south-central Texas. Front. Ecol. Environ. 4, 238–243. https://doi.org/10.1890/1540-9295(2006)004[0238:Evotpc]2.0.Co;2 (2006).
Article  Google Scholar 

4.
Boyles, J. G., Cryan, P. M., McCracken, G. F. & Kunz, T. H. Conservation. Economic importance of bats in agriculture. Science 332, 41–42. https://doi.org/10.1126/science.1201366 (2011).
Article  PubMed  ADS  Google Scholar 

5.
Naylor, R. L. & Ehrlich, P. R. In Nature’s Services: Societal Dependence on Natural Ecosystems (ed. Daily, G. C.) 151–174 (Island Press, New York, 1997).
Google Scholar 

6.
Losey, J. E. & Vaughan, M. The economic value of ecological services provided by insects. Bioscience 56, 311–323. https://doi.org/10.1641/0006-3568(2006)56[311:Tevoes]2.0.Co;2 (2006).
Article  Google Scholar 

7.
Power, A. G. Ecosystem services and agriculture: Tradeoffs and synergies. Philos. Trans. R. Soc. Lond. B 365, 2959–2971. https://doi.org/10.1098/rstb.2010.0143 (2010).
Article  Google Scholar 

8.
Maine, J. J. & Boyles, J. G. Bats initiate vital agroecological interactions in corn. Proc. Natl. Acad. Sci. U.S.A. 112, 12438–12443. https://doi.org/10.1073/pnas.1505413112 (2015).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

9.
Tremblay, A., Mineau, P. & Stewart, R. K. Effects of bird predation on some pest insect populations in corn. Agric. Ecosyst. Environ. 83, 143–152. https://doi.org/10.1016/S0167-8809(00)00247-4 (2001).
Article  Google Scholar 

10.
Van Bael, S. A. et al. Birds as predators in tropical agroforestry systems. Ecology 89, 928–934 (2008).
Article  Google Scholar 

11.
Grass, I., Lehmann, K., Thies, C. & Tscharntke, T. Insectivorous birds disrupt biological control of cereal aphids. Ecology 98, 1583–1590. https://doi.org/10.1002/ecy.1814 (2017).
Article  PubMed  Google Scholar 

12.
Karp, D. S. et al. Forest bolsters bird abundance, pest control and coffee yield. Ecol. Lett. 16, 1339–1347. https://doi.org/10.1111/ele.12173 (2013).
Article  PubMed  Google Scholar 

13.
Maas, B. et al. Bird and bat predation services in tropical forests and agroforestry landscapes. Biol. Rev. Camb. Philos. Soc. 91, 1081–1101. https://doi.org/10.1111/brv.12211 (2016).
Article  PubMed  Google Scholar 

14.
Cohen, Y., Bar-David, S., Nielsen, M., Bohmann, K. & Korine, C. An appetite for pests: Synanthropic insectivorous bats exploit cotton pest irruptions and consume various deleterious arthropods. Mol. Ecol. 29, 1185–1198. https://doi.org/10.1111/mec.15393 (2020).
Article  PubMed  Google Scholar 

15.
Chaplin-Kramer, R., de Valpine, P., Mills, N. J. & Kremen, C. Detecting pest control services across spatial and temporal scales. Agric. Ecosyst. Environ. 181, 206–212. https://doi.org/10.1016/j.agee.2013.10.007 (2013).
Article  Google Scholar 

16.
Speakman, J. R. & Thomas, D. W. In Bat ecology (eds Kunz, T. H. & Fenton, M. B.) 430–490 (University of Chicago Press, Chicago, 2003).
Google Scholar 

17.
Norberg, U. M. Avian Energetics and Nutritional Ecology 199–249 (Springer, Berlin, 1996).
Google Scholar 

18.
Nyffeler, M., Şekercioğlu, Ç. H. & Whelan, C. J. Insectivorous birds consume an estimated 400–500 million tons of prey annually. Sci. Nat. 105, 47. https://doi.org/10.1007/s00114-018-1571-z (2018).
CAS  Article  Google Scholar 

19.
Sekercioglu, C. H. Increasing awareness of avian ecological function. Trends Ecol. Evol. 21, 464–471. https://doi.org/10.1016/j.tree.2006.05.007 (2006).
Article  PubMed  Google Scholar 

20.
Mols, C. M. M. & Visser, M. E. Great tits can reduce caterpillar damage in apple orchards. J. Appl. Ecol. 39, 888–899. https://doi.org/10.1046/j.1365-2664.2002.00761.x (2002).
Article  Google Scholar 

21.
Van Bael, S. A., Bichier, P. & Greenberg, R. Bird predation on insects reduces damage to the foliage of cocoa trees (Theobroma cacao) in western Panama. J. Trop. Ecol. 23, 715–719. https://doi.org/10.1017/s0266467407004440 (2007).
Article  Google Scholar 

22.
Federico, P. et al. Brazilian free-tailed bats as insect pest regulators in transgenic and conventional cotton crops. Ecol. Appl. 18, 826–837. https://doi.org/10.1890/07-0556.1 (2008).
Article  PubMed  Google Scholar 

23.
McCracken, G. F. et al. Bats track and exploit changes in insect pest populations. PLoS ONE 7, e43839. https://doi.org/10.1371/journal.pone.0043839 (2012).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

24.
Kolkert, H., Andrew, R., Smith, R., Rader, R. & Reid, N. Insectivorous bats selectively source moths and eat mostly pest insects on dryland and irrigated cotton farms. Ecol. Evol. 10, 371–388. https://doi.org/10.1002/ece3.5901 (2020).
Article  PubMed  Google Scholar 

25.
Maas, B., Clough, Y. & Tscharntke, T. Bats and birds increase crop yield in tropical agroforestry landscapes. Ecol. Lett. 16, 1480–1487. https://doi.org/10.1111/ele.12194 (2013).
Article  PubMed  Google Scholar 

26.
Kalka, M. B., Smith, A. R. & Kalko, E. K. Bats limit arthropods and herbivory in a tropical forest. Science 320, 71. https://doi.org/10.1126/science.1153352 (2008).
CAS  Article  PubMed  ADS  Google Scholar 

27.
Williams-Guillen, K., Perfecto, I. & Vandermeer, J. Bats limit insects in a neotropical agroforestry system. Science 320, 70. https://doi.org/10.1126/science.1152944 (2008).
CAS  Article  PubMed  ADS  Google Scholar 

28.
Kunz, T. H., de Torrez, E. B., Bauer, D., Lobova, T. & Fleming, T. H. In Year in Ecology and Conservation Biology (eds Ostfeld, R. S. & Schlesinger, W. H.) 1–38 (New York Academy of Sciences, New York, 2011).
Google Scholar 

29.
Taylor, P. J., Grass, I., Alberts, A. J., Joubert, E. & Tscharntke, T. Economic value of bat predation services: A review and new estimates from macadamia orchards. Ecosyst. Serv. 30, 372–381. https://doi.org/10.1016/j.ecoser.2017.11.015 (2018).
Article  Google Scholar 

30.
Redlich, S., Martin Emily, A. & Steffan-Dewenter, I. Landscape-level crop diversity benefits biological pest control. J. Appl. Ecol. https://doi.org/10.1111/1365-2664.13126 (2018).
Article  Google Scholar 

31.
Hooks, C., Pandey, R. R., & Johnson, M. W. Unlikely guardians of cropping systems: Can birds and spiders protect broccoli from caterpillar pests? Insect Pests (2007).

32.
Martin, E. A., Reineking, B., Seo, B. & Steffan-Dewenter, I. Natural enemy interactions constrain pest control in complex agricultural landscapes. Proc. Natl. Acad. Sci. U.S.A. 110, 5534–5539. https://doi.org/10.1073/pnas.1215725110 (2013).
CAS  Article  PubMed  PubMed Central  ADS  Google Scholar 

33.
Karp, D. S. & Daily, G. C. Cascading effects of insectivorous birds and bats in tropical coffee plantations. Ecology 95, 1065–1074. https://doi.org/10.1890/13-1012.1 (2014).
Article  PubMed  Google Scholar 

34.
Barbaro, L. et al. Avian pest control in vineyards is driven by interactions between bird functional diversity and landscape heterogeneity. J. Appl. Ecol. 54, 500–508. https://doi.org/10.1111/1365-2664.12740 (2017).
Article  Google Scholar 

35.
Rey Benayas, J. M., Meltzer, J., de las Heras-Bravo, D. & Cayuela, L. Potential of pest regulation by insectivorous birds in Mediterranean woody crops. PLoS ONE 12, 15. https://doi.org/10.1371/journal.pone.0180702 (2017).
CAS  Article  Google Scholar 

36.
Morrison, E. B. & Lindell, C. A. Birds and bats reduce insect biomass and leaf damage in tropical forest restoration sites. Ecol. Appl. 22, 1526–1534 (2012).
Article  Google Scholar 

37.
Ndanganga, P. K., Njoroge, J. B. M. & Vickery, J. Quantifying the contribution of birds to the control of arthropod pests on kale, Brassica oleracea acephala, a key crop in East African highland farmland. Int. J. Pest Manage. https://doi.org/10.1080/09670874.2013.820005 (2013).
Article  Google Scholar 

38.
Tschumi, M., Ekroos, J., Hjort, C., Smith, H. G. & Birkhofer, K. Rodents, not birds, dominate predation-related ecosystem services and disservices in vertebrate communities of agricultural landscapes. Oecologia 188, 863–873. https://doi.org/10.1007/s00442-018-4242-z (2018).
Article  PubMed  PubMed Central  ADS  Google Scholar 

39.
Elkinton, J. S., Liebhold, A. M. & Muzika, R.-M. Effects of alternative prey on predation by small mammals on gypsy moth pupae. Popul. Ecol. 46, 171–178. https://doi.org/10.1007/s10144-004-0175-y (2004).
Article  Google Scholar 

40.
Dyrcz, A. Breeding biology and behaviour of the willie wagtail (Rhipidura leucophrys) in the madang region Papua New Guinea. Emu 94, 17–26. https://doi.org/10.1071/MU9940017 (1994).
Article  Google Scholar 

41.
Adriano, S. & Calver, M. C. Diet of breeding willie wagtails (Rhipidura leucophrys) in suburban Western Australia. Emu 95, 138–141. https://doi.org/10.1071/MU9950138 (1995).
Article  Google Scholar 

42.
Razeng, E. & Watson, D. M. What do declining woodland birds eat? A synthesis of dietary records. Emu 112, 149–156. https://doi.org/10.1071/MU11099 (2012).
Article  Google Scholar 

43.
Brandl, R., Kristín, A. & Leisler, B. Dietary niche breadth in a local community of passerine birds: An analysis using phylogenetic contrasts. Oecologia 98, 109–116. https://doi.org/10.1007/bf00326096 (1994).
CAS  Article  PubMed  ADS  Google Scholar 

44.
Kaplan, G. Australian Magpie: Biology and Behaviour of an Unusual Songbird (CSIRO Publishing, Clayton, 2019).
Google Scholar 

45.
Puckett, H. L., Brandle, J. R. & Johnson, R. J. Avian foraging patterns in crop field edges adjacent to woody habitat. Agric. Ecosyst. Environ. 131, 9–15 (2009).
Article  Google Scholar 

46.
Best, L. B., Whitmore, R. C. & Booth, G. M. Use of cornfields by birds during the breeding season: the importance of edge habitat. Am. Midl. Nat. 123, 84–99. https://doi.org/10.2307/2425762 (1990).
Article  Google Scholar 

47.
Hansen, N. A., Sato, C. F., Michael, D. R., Lindenmayer, D. B. & Driscoll, D. A. Predation risk for reptiles is highest at remnant edges in agricultural landscapes. J. Appl. Ecol. 56, 31–43. https://doi.org/10.1111/1365-2664.13269 (2019).
Article  Google Scholar 

48.
Storch, I., Woitke, E. & Krieger, S. Landscape-scale edge effect in predation risk in forest-farmland mosaics of Central Europe. Landsc. Ecol. 20, 927–940. https://doi.org/10.1007/s10980-005-7005-2 (2005).
Article  Google Scholar 

49.
Douglas, D. J. T., Vickery, J. A. & Benton, T. G. Improving the value of field margins as foraging habitat for farmland birds. J. Appl. Ecol. 46, 353–362. https://doi.org/10.1111/j.1365-2664.2009.01613.x (2009).
Article  Google Scholar 

50.
Stephens, D. W. & Krebs, J. R. Foraging Theory (University Press, 1986).
Google Scholar 

51.
Denzinger, A. & Schnitzler, H.-U. Bat guilds, a concept to classify the highly diverse foraging and echolocation behaviors of microchiropteran bats. Front. Physiol. 4, 164. https://doi.org/10.3389/fphys.2013.00164 (2013).
Article  PubMed  PubMed Central  Google Scholar 

52.
Schnitzler, H.-U. & Kalko, E. K. V. Echolocation by insect-eating bats. Vol. 51 (SPIE, 2001).

53.
Neuweiler, G. Foraging, echolocation and audition in bats. Naturwissenschaften 71, 446–455. https://doi.org/10.1007/BF00455897 (1984).
Article  ADS  Google Scholar 

54.
Fenton, M. B. The foraging behaviour and ecology of animal-eating bats. Can. J. Zool. 68, 411–422. https://doi.org/10.1139/z90-061 (1990).
Article  Google Scholar 

55.
Jantzen, M. K. & Fenton, M. B. The depth of edge influence among insectivorous bats at forest–field interfaces. Can. J. Zool. 91, 287–292. https://doi.org/10.1139/cjz-2012-0282 (2013).
Article  Google Scholar 

56.
Estur, G. Cotton Exporter’s Guide. (International Trade Centre UNCTAD/WTO, 2007).

57.
RBG In State of the World’s Plants 2017 (ed. Willis, K.) 64–71 (Board of Trustees of the Royal Botanic Gardens, Kew, 2017).
Google Scholar 

58.
Rencken, I. An Investigation of the Importance of Native and Non-Crop Vegetation to Beneficial Generalist Predators in Australian Cotton Agroecosystems PhD thesis, University of New England (2006).

59.
Holloway, J. C., Furlong, M. J. & Bowden, P. I. Management of beneficial invertebrates and their potential role in integrated pest management for Australian grain systems. Aust. J. Exp. Agric. 48, 1531–1542. https://doi.org/10.1071/EA07424 (2008).
Article  Google Scholar 

60.
Schellhorn, N. A., Bianchi, F. J. & Hsu, C. L. Movement of entomophagous arthropods in agricultural landscapes: Links to pest suppression. Annu. Rev. Entomol. 59, 559–581. https://doi.org/10.1146/annurev-ento-011613-161952 (2014).
CAS  Article  PubMed  Google Scholar 

61.
Whitehouse, M. E. A., Wilson, L. J. & Fitt, G. P. A comparison of arthropod communities in transgenic Bt and conventional cotton in Australia. Environ. Entomol. 34, 1224–1241 (2005).
Article  Google Scholar 

62.
Smith, R., Reid, J., Scott-Morales, L., Green, S. & Reid, N. A baseline survey of birds in native vegetation on cotton farms in inland eastern Australia. Wildl. Res. 46, 304–316. https://doi.org/10.1071/WR18038 (2019).
Article  Google Scholar 

63.
Ford, G. & Thomson, N. Birds on Cotton Farms: A Guide to Common Species and Habitat Management (Cotton Catchment Communities CRC, Boca Raton, 2006).
Google Scholar 

64.
Whelan, C. J., Wenny, D. G. & Marquis, R. J. In Year in Ecology and Conservation Biology (eds Ostfeld, R. S. & Schlesinger, W. H.) 25–60 (Annals of the New York Academy of Sciences, New York, 2008).
Google Scholar 

65.
Rodríguez, A., Andrén, H. & Jansson, G. Habitat-mediated predation risk and decision making of small birds at forest edges. Oikos 95, 383–396 (2001).
Article  Google Scholar 

66.
Sekercioglu, C. H. Bird functional diversity and ecosystem services in tropical forests, agroforests and agricultural areas. J. Ornithol. 153, 153–161. https://doi.org/10.1007/s10336-012-0869-4 (2012).
Article  Google Scholar 

67.
Greenop, A., Woodcock, B. A., Wilby, A., Cook, S. M. & Pywell, R. F. Functional diversity positively affects prey suppression by invertebrate predators: a meta-analysis. Ecology 99, 1771–1782. https://doi.org/10.1002/ecy.2378 (2018).
Article  PubMed  PubMed Central  Google Scholar 

68.
Nelson, J. J. & Gillam, E. H. Selection of foraging habitat by female little brown bats (Myotis lucifugus). J. Mammal. 98, 222–231. https://doi.org/10.1093/jmammal/gyw181 (2016).
Article  Google Scholar 

69.
Rohner, C. & Krebs, C. J. Owl predation on snowshoe hares: Consequences of antipredator behaviour. Oecologia 108, 303–310. https://doi.org/10.1007/bf00334655 (1996).
Article  PubMed  ADS  Google Scholar 

70.
Rockwell, C., Gabriel, P. O. & Black, J. M. Bolder, older, and selective: Factors of individual-specific foraging behaviors in Steller’s jays. Behav. Ecol. 23, 676–683. https://doi.org/10.1093/beheco/ars015 (2012).
Article  Google Scholar 

71.
Krebs, J. R. In Perspectives in Ethology (eds Bateson, P. P. G. & Klopfer, P. H.) 73–111 (Springer US, Berlin, 1973).
Google Scholar 

72.
Lövei, G. L. & Ferrante, M. A review of the sentinel prey method as a way of quantifying invertebrate predation under field conditions. Insect Sci. 24, 528–542. https://doi.org/10.1111/1744-7917.12405 (2017).
Article  PubMed  Google Scholar 

73.
Nagy, R. K., Schellhorn, N. A. & Zalucki, M. P. Fresh, frozen or fake: A comparison of predation rates measured by various types of sentinel prey. J. Appl. Entomol. 144, 407–416. https://doi.org/10.1111/jen.12745 (2020).
Article  Google Scholar 

74.
Ravzanaadii, N., Kim, S.-H., Choi, W. H., Seong-Jin, H. & Kim, N. J. Nutritional value of mealworm, tenebrio molitor as food source. Int. J. Ind. Ergon. 25, 93–98 (2012).
Google Scholar 

75.
Barbaro, L., Giffard, B., Charbonnier, Y., van Halder, I. & Brockerhoff, E. G. Bird functional diversity enhances insectivory at forest edges: A transcontinental experiment. Divers. Distrib. 20, 149–159. https://doi.org/10.1111/ddi.12132 (2014).
Article  Google Scholar 

76.
Environment Australia. Revision of the Interim Biogeographic Regionalisation of Australia (IBRA) and Development of Version 5.0: Summary Report. (Department of Environment and Heritage, Canberra, 2000).

77.
OEH. NSW (Mitchell) Landscapes. Vol. version 3.1 (State of New South Wales and Office of Environment and Heritage 2002).

78.
Keith, D. A. Ocean Shores to Desert Dunes: The Native Vegetation of NEW South Wales and the ACT (Department of Environment and Conservation, 2004).
Google Scholar 

79.
R Core Team, R: A language and environment for statistical computing. (R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/, 2018).

80.
Brooks, M. E. et al. glmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling. R J. 9, 400. https://doi.org/10.3929/ethz-b-000240890 (2017).
Article  Google Scholar 

81.
Lüdecke, D. ggeffects: Create Tidy Data Frames of Marginal Effects for ‘ggplot’ from Model Outputs (v0.16.0). https://CRAN.R-project.org/package=ggeffects. (2017).

82.
Hartig, F. DHARMa: Residual Diagnostics for Hierarchical (Multi-level/mixed) Regression Models (v0.3.3). https://CRAN.R-project.org/package=DHARMa (2017).

83.
Lenth, R., Singmann, H., Love, J., Buerkner, P. & Herve, M. Estimated Marginal Means, aka Least-Squares Means (v1.5.3). https://www.rdocumentation.org/packages/emmeans (2019).

84.
Szumilas, M. Explaining odds ratios. J. Can. Acad. Child Adolesc. Psychiatry 19, 227–229 (2010).
Article  Google Scholar  More

This research presents HD-ClvR, which is a combination of three gene drives: homing, cleave-and-rescue and daisyfield. Our modelling indicates that HD-ClvR overcomes an important trade-off in current homing gene drive designs: the trade-off between resistance allele formation and gene drive efficiency. This strategy benefits from the efficiency of a homing gene drive and the evolutionary stability of cleave-and-rescue gene drive. Due to the inclusion of a daisyfield system, HD-ClvR is self-limiting and can be controlled by supplementation of gene drive animals.
HD-ClvR compared to other gene drives
Over recent years, many different gene drives have been published and developments have been geared towards both efficiency and safety38. An ongoing issue has been the development of resistance alleles. For CRISPR-based homing gene drive there are two fundamental approaches to combat resistance allele formation: careful gRNA targeting and gRNA multiplexing. When a gRNA targets a conserved sequence in a gene, resistance alleles are likely to disrupt gene function through NHEJ repair and will therefore reduce fitness39. Recently, population suppression was already shown to work with a carefully targeted homing gene drive in contained mosquito populations39, however, current data suggests that homing might be less efficient in mammals than in insects14. A recent paper has proposed the concept of ‘tethered homing gene drive’, which combines a threshold-dependent underdominance gene drive with a homing gene drive for improved suppression capabilities40. We use this concept in a different manner in HD-ClvR, by relying on a daisyfield rather than threshold-dependence for self-limitation. Very recently, two new papers have proposed a gene drive similar to HD-ClvR, but intented for population modification instead of suppression41,42. These studies also combine homing and cleave-and-rescue principles to combat resistance alleles and their modifications are able to persist stably in cage experiments, which is promising for HD-ClvR.
In addition to targeting conserved sequences, when gRNA multiplexing, resistant allele allele formation is reduced because multiple sites are targeted simultaneously. For homing gene drives, multiplexing has been shown to reduce homing efficiency when more than two gRNAs are used28. In contrast, cleave-and-rescue gene drives do not have this problem, as they do not use homing and can therefore multiplex gRNAs without any efficiency costs. HD-ClvR separates the elimination of resistance alleles and homing efficiency, and therefore gRNAs can be optimised for both goals separately.
To date, most gene drive research has focused on improving the efficiency, however, equally important is the development of strategies that allow for containment, or even reversibility, of the gene drives29,43. For contained gene drives, density dependence is often used, which requires large numbers of gene drive individuals to be released into a target population to spread44. Therefore, non-target populations are unlikely to be affected by this type of gene drive. However, a large single release of gene drive individuals can put significant pressure on the local ecosystem, and if a population is already at carrying capacity, it may lead to starvation or mass migration of the population. In contrast, HD-ClvR uses ongoing input in the form of gene drive animals to control the extent of population suppression and contain spread, while the total amount of gene drive animals necessary for release is similar to threshold-dependent gene drives. Therefore, the use of HD-ClvR seems more feasible than threshold-dependent gene drives. Although self-limitation comes with increased cost and labour relative to unlimited gene drives, we believe this is justified by the control and safety of HD-ClvR.
As stated above, the initial introduction frequency for a standard cleave-and-rescue gene drive in our randomly mating model was increased 10-fold over the other homing-based strategies. This increase is necessary due to the significant cost to the reproduction rate that is incurred when using a standard cleave-and-rescue gene drive. On average, cleave-and-rescue animals will produce 50% less offspring than wild-type animals21,24. This significantly slows the spread of the gene drive and due to density dependent dynamics, requires large initial releases of cleave-and-rescue animals for population suppression. With a homing-cleave-and-rescue drive, more offspring inherit the drive and there is less cost to the reproduction rate. Effectively, for homing-cleave-and-rescue, the reproduction rate of gene drive individuals is equal to the homing efficiency (plus half of the homing failure rate, where the gene drive is inherited by chance), which so far has been shown to range from 0.7 to 1 in different organisms14,39,45.
Supplementation
As animal supplementation is a critical component of HD-ClvR, our modelling investigated how daisyfield size and the level and placement of supplemented HD-ClvR animals effects efficiency and safety of population suppression. Optimisation of these parameters can significantly reduce cost and labour, as well as reduce the risk of unwanted impacts on non-target populations. We modelled our supplementation as a percentage of the total population size, therefore the number of individuals needed for supplementation increases linearly with population size. We also want to minimise the risk of non-target populations being impacted by the gene drive, and therefore, there is a trade-off between safety (size of the daisyfield) and cost and labour (level of supplementation required).
The least number of daisy elements that can suppress the population with a realistic level of supplementation, but does not cause any serious issues in non-target populations, should be objectively established through an in-depth risk assessment process. In a larger population however, the spread is slower than in a small one. Therefore, for improved safety and efficiency, gene drives are best applied in small sub-populations separately. The impact of a single introduction, such as a rogue deployment or migration, depends on the population size. The smaller the population, the bigger the impact. This it is a concern when the target population is much larger than the non-target population, but this is not the case for invasive UK grey squirrels and many other invasive species.
The appropriate daisyfield size also depends on the rate of NHEJ ((P_n)) of the gene drive system; the higher the ((P_n)), the more embryonic lethal offspring will arise and the sooner daisyfield burns out. To choose a safe number of daisy elements, we also need an estimate of how many animals a rogue party could obtain, potential breed and add into a non-target population for their own benefit. Overall, each target population and prospective gene drive strategy needs to be considered on a case-by-case basis and include an in-depth multidisciplinary risk assessment process.
When we consider the spatial aspects of a HD-ClvR supplementation programme, the picture becomes more complex. A key factor is the supplementation location of individuals. Obviously, supplementing individuals in a location where the population has already been suppressed will be ineffective. Therefore, different placement strategies can be adopted to keep placing individuals in a relevant area. A monitoring system where not only the size of the population is known, but also the location can significantly help HD-ClvR continue spreading and suppress a targeted population.
In this study, we modelled HD-ClvR using five different supplementation placement strategies in grey squirrel. These were: supplementation at the mean of population location, the mode of population location, randomly, randomly in 10 groups, and in a moving front (Fig. 6a). With supplementation at the mean of the population location, supplementation started in the middle of the population. After a few generations, a gap appears in the middle due to local suppression. The mean of the populations location still lies in the middle, as can be seen in Fig. 6c at 20 generations. Therefore, supplementation is not effective until the population is also suppressed in another location, thereby shifting the mean. Additionally, when there is a single large patch of the population left and additional smaller clusters, supplementation in the middle of the large patch allows the smaller clusters to recover, as can be seen in Fig. 6c after 64 generations.
With supplementation at the mode of the population location, we supplement in a location where there are many individuals. This placement strategy avoids the problem of supplementing in a location without individuals, either in a doughnut-like spatial population structure or in a multi-patch population. However, this placement strategy still allows small patches to form and recover. Supplementation at a random location theoretically means that supplementation happens uniformly, but in reality, this is not the case. Initially HD-ClvR spreads in multiple locations, but after the population is suppressed in certain regions, supplementation in those regions becomes ineffective. Therefore, at a later stage of population suppression this placement scheme becomes increasingly ineffective.
Supplementation at random locations is more effective when they are broken up into multiple groups (ten in our model). The gene drive spreads in many locations initially like the random single location placement scheme. After significant suppression of the population some but not all of the 10 groups supplemented are at ineffective locations. The groups that are placed at relevant locations are enough to keep the gene drive spreading. In our model supplementation in groups at random locations gets close to the speed at which a gene drive spreads in a non-spatial model.
The moving front placement scheme is very effective initially, as the gene drive spreads uniformly across the front. In this case, supplementation keeps ahead of where the populations is being suppressed. This placement strategy allows the population to recover behind the moving front after effective initial spread and near-complete suppression. To improve efficiency of the moving front strategy, it may be beneficial to include random supplementation behind the moving front to prevent animals from re-establishing.
Finally, in our spatial model, it was evident that there is more uncertainty in levels of population suppression than a randomly mating model leads us to believe. As can be seen in Fig. 6b, the 95% quantiles are broader than the quantiles in Fig. 3. Therefore, we conclude that to tailor the amount of supplementation, it is vital to closely monitor a population where a gene drive is used.
Assumptions and future work
Our model works under the following six assumptions. First, our model excludes some complexities of the optimal number of gRNAs for homing. Although our model suggests that multiplexing gRNAs for both the homing and cleave-and-rescue gene drives is most effective, a recent study using a more complex model and in vivo data shows that the optimal number of gRNAs to use for homing in Drosphilia melanogaster is two. They report a decrease in homing efficiency with more than two gRNAs due to reduced homology and Cas nuclease saturation28. Therefore, our gene drive with four gRNAs for both homing and cleave-and-rescue will likely be less efficient in such a complex model. We suggest using two homing gRNAs and four cleave- and-rescue gRNAs is likely most efficient, while still eliminating all resistance alleles28. It would be prudent to analyse our gene drive in this complex model as well to get a definitive estimate, as Cas saturation is thought to have an influence on gene drive efficiency when multiplexing is used28.
Second, we assumed there was no embryonic Cas-gRNA expression. Embryonic Cas-gRNA expression might be problematic as it leads to resistance allele formation and can interfere with the cleave-and-rescue mechanism by cleaving alleles from the wildtype parent. As our gene drive eliminates resistance alleles, embryonic Cas-gRNA expression may not inhibit spread, depending on the rate. Additionally, if the embryonic Cas-gRNA expression turns out to be more common in grey squirrel or other species, the cleave-and-rescue part of the gene drive can be harnessed with a double rescue mechanism to overcome this issue, as reported by Champer et al.24.
Third, we did not take other types of resistance alleles into account such as mutations rendering the CRISPR-Cas non-functional. As this is a universal assumption in gene drive research, we will have to await multigenerational studies to see if this is problematic.
Fourth, HD-ClvR has not been tested in vivo, which is our next step. The two recent papers testing a gene drive similar to HD-ClvR for population modification have performed in vivo tests in Drosophila melanogaster which showed very efficient conversion rates41. Proof-of-concept testing of HD-ClvR would likely initially occur in D. melanogaster and mouse models before progressing to squirrel studies. Recent reports have shown that the VASA promoter for Cas expression in homing gene drives is not optimal and further investigation to identify a meiosis-specific germline promoter is needed15. Furthermore, the integration of many daisies in a squirrel genome will be a molecular challenge and is a feat which has not yet been reported on in any species. This task could be achieved using either a random integration strategy, such as lentiviruses46 or a targeted integration strategy that exploits neutral repetitive sequences in the genome as target sites32. Also, non-model species might be difficult to genetically engineer, although grey squirrel embryology will likely follow the extensive knowledge on rodent and farmed animal embryology, and similar reagents and equipment could be used. An important consideration when engineering gene drive is that the modified animals maintain enough wild vigour to survive and breed in a wild population. Promising technologies for generating gene drive harbouring mammals with as little intervention as possible include in situ delivery of CRISPR reagents to the oviduct47.
Fifth, for our spatial modelling, we assumed that an estimation of population size could be made every year, although there is a significant amount of room for error in this estimate. Additionally, for some of our placement schemes, we assumed an accurate estimate of population location. As the random placement in groups scheme turned out most effective, this is not a problem so much as further potential for improvement. Another direction for future spatial work is the modelling of real landscapes, which are more complex than what we modelled in this study48. In complex landscapes, it might be that gene drive spread is slower or even regionally confined in some situations. Additionally, there might be spatial dynamics to gene drives in general such as ’chasing’, which is the perpetual escaping and chasing of wildtype and gene drive animals34. Further efforts are necessary to create a more realistic spatial model before we can consider using a gene drive.
A final consideration is that the ecological services the grey squirrel and other invasive species provide are largely unchartered. Ecologists need to investigate the ecological services that an invasive species performs and how an abrupt suppression of this invasive population might impact the ecosystem as a whole. We need to consider other restorative measures such as reintroducing native species to fragmented habitats, amongst other ecological interventions49. From a regulatory perspective, there is no tested legislative framework for the release of gene drive organisms; and with regard to our test animal it is currently illegal to breed grey squirrels in the UK. Developing these legislative frameworks alongside gene drive research is important. More importantly, the UK needs to continue to broaden public engagement and see whether the public is receptive to the deployment of gene drive technology in parallel to a financial overview of how much it would cost to apply gene drives reflecting our predicted need for supplementation.
Summary
HD-ClvR offers an efficient, self-limiting, and controllable gene drive strategy. We show that in the spatial model, complete population suppression is achieved approximately 5 years later than in the randomly mating population model. We then explored how the placement of supplemented animals could impact population suppression. Our results show that spatial dynamics of supplementation placement are not prohibitive to the spread of the gene drive, but that in fact, with an optimised strategy, spread at a rate equal to randomly mating population can be achieved. In our models, we have shown that grey squirrels have a spatial life history which facilitates the spread of a gene drive. Therefore, gene drives could be a valuable tool in the conservation toolbox. More