Ecology

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

Extensive hydrodynamic model validation
To provide a comprehensive validation of the modeled surge-tide-wave dynamics during Sandy, this study made use of all available field data from an extensive network of coastal water level gauges and wave buoys in the study region, as well as hundreds of HWMs and many rapid deployment surge sensors (SI Fig. S1). The good agreements between simulated and observed water levels and waves were quantified in terms of the root-mean-square error (RMSE) and correlation coefficient (CORR).
Storm tide
Available data included those from hundreds of permanent and temporary surge sensors from NOAA (SI Fig. S2), Hudson River Environmental Conditions Observing System (HRECOS), and USGS. Statistics (SI Table S1) showed excellent agreement between the time series of simulated and measured storm tide data. The averaged RMSE at USGS temporary sites was 0.20 m and the averaged CORR was 0.94. CH3D-SSMS successfully reproduced the surge and tides with high confidence in not only the open ocean but also the complex estuarine system during Sandy (SI Fig. S3) with a maximum coastal storm tide at NY Bight of 3.36 m and NJ coast of 3.27 m.
High water marks and inundation
A total of 526 out of 653 independent HWM locations were located within the study region, most of which clustered in NJ and NY coastal areas. 83.7% (440 out of 526) of HWMs were captured by the CH3D-SSMS model grid and compared with the surveyed values. The model results had 0.33 m RMSE and 0.87 CORR (SI Fig. S4). When only “good” and “excellent” HWMs (rated by USGS) were used, the RMSE dropped to 0.30 m, and CORR increased to 0.90. The noticeable data disagreements (Fig. SI 4) were caused by the inconsistency of surveyed data.
Wave
During Sandy, four wave-buoys within the domain recorded the wave data: two at the apex of NY Bight, and another two in Long Island Sound (LIS). The significant wave height (H_{sig}) and peak wave period (T_{p}) simulated by the Simulating Waves Nearshore (SWAN) model in the CH3D-SSMS were compared with measured data and Wave Watch III (WW3) operational run results (SI Fig. S5)25,26. SWAN more accurately captured the evolution of (H_{sig}) in NY Bight and LIS. Maximum significant wave height over land was as high as 2.17 m at NY Bight and 2.60 m along NJ coast.
Impact of wetland on surge and wave during super storm sandy
To estimate the value of the coastal wetlands in reducing flood, we calculated the following four metrics for inundation: the Average Inundation Height ((AIH)), the Maximum Inundation Height ((MIH)), the Total Inundation Area ((TIA)), and the Total Inundation Volume ((TIV)) with and without wetlands. The (TIA) and (TIV) are defined as

$$ TIA = iint_{{{text{Landward}};{text{area}}}} {dxdy} , $$
(1)

$$ TIV = iint_{{{text{Landward}};{text{area}}}} [H_{max} left( {x,y} right) – H_{0} left( {x,y} right)]dxdy, $$
(2)

where (H_{max} left( {x,y} right)) and (H_{0} left( {x,y} right)) are the maximum water level and the land elevation at land cells (left( {x,y} right)), respectively11,12. The wave analysis was carried out by calculating the average wave height ((AWH)), the maximum wave height ((MWH)), and the total wave energy ((TWE)) which is defined as

$$ TWE = iint_{{{text{Landward}};{text{area}}}} [frac{1}{8}rho_{w} g{ }left( {H_{{rms,max{ }}} } right)^{2} ]dxdy, $$
(3)

where (H_{rms max}) is the maximum root-mean-square wave height over the flooded land. The wave energy, instead of wave height, is more directly related to the wave-induced structure loss.
We define the relative inundation reduction ((RIR)) as the difference in (TIV) (value without wetland minus value with wetland), divided by the (TIV) with wetland. The relative wave reduction ((RWR)) is defined accordingly using the (TWE). The inundation and wave analysis were carried out at the regional level (SI Table S2) and zip-code resolution (Figs. 1, S6, S7).
Figure 1

Zip-code resolution wetland’s effect on (TIV) and (TWE) during Sandy in 2012. Map showing zip-code resolution avoidance in (A) (TIV) and (B) (TWE) during Sandy without wetlands, as a percentage of those for the with-wetland scenario. Dark red values show zip-code with the most wetland benefit while dark blue areas have the least wetland benefit. Negative values indicate that the presence of wetland would increase (TIV)/(TWE) and positive values indicate that wetland would lower (TIV)/(TWE). The map is produced using ESRI ArcGIS Pro 2.7 (https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview).

Full size image

If the wetlands were absent during Sandy, the (RIR) of the entire model domain would have been 4% and the (RWR) would have been 19%. Breaking the entire model domain into 6 regions (Table S2): South NJ (SNJ), Central NJ (CNJ), North NJ (NNJ), CT, Long Island (LI), and Mainland NY (MNY), it was found that the (RIR) during sandy was low ( 50%) (RWR), while LI. CNJ and SNJ had significant (~ 25–50%) (RWR). The regions with the highest (RIR) and (RWR) were closest to the storm landfall location.
Benefit of wetland on surge and wave for the 1% annual chance flood
The maximum 1% annual chance maximum flood elevation (Fig. S9) was 5.15 m for NY Bight while 5.34 m for the NJ coast. The 1% annual chance maximum wave height (Fig. S9) overland is 2.24 m at NY Bight and 1.90 m at NJ coast. All the regions had moderate (RIR) except for SNJ which had a significant (RIR) due to highest (25.4%) wetland cover (see Table S2) and NNJ experienced significant (RWR) while the remaining regions had moderate values. CT ranked 2nd in (RIR ) although it has less wetland cover than CNJ because the mostly woody wetlands in CT are more effective in buffering storm surge than the marshes in NJ and NY. On the other hand, CT ranked 3rd in (RWR) due to the blocking of offshore wave energy by LI. (RIR) and (RWR) are found to be functions of storm characteristics, wetland type, and cover, and local conditions. Fig. S10 shows the percent wetland cover, RIR (relative TIV reduction), and RWR (relative wave energy reduction) in six regions (New York, North New Jersey, Long Island, Connecticut, Central New Jersey, and South New Jersey) during 1% annual chance events. As the wetland cover increases from less than 5% (NY, NNJ, and LI) to more than 10% in CNJ and SNJ, RIR and RWR generally increase, showing the increasing role of wetland in reducing inundation and wave. Relative reduction in inundation and wave energy are modest, between 10 and 30%. NNJ has properties behind the relatively sparse marsh, followed by woody wetland which protects properties behind them. Connecticut has less wetland than Central Jersey, but the mostly woody wetland is more effective in reducing flood and wave.
Benefit of wetlands on reducing residential structure loss
The monetary loss of residential structures was estimated using the simulated inundation and wave results while employing damage functions from the United States Army Corps of Engineer (USACE) North Atlantic Comprehensive Coastal Study (NACCS) and was validated using the NFIP building loss payouts aggregated by zip-code21,24. Direct simulation of wave-induced damage requires understanding and calculation of wave loads on structures using a depth- and phase-resolving model, which is beyond the capability of the models used in this study. Therefore, in this paper, we did not directly simulate wave-induced damage, but are accounting for wave-induced change in total water level which results in increase in estimated damage based on depth-damage functions. Overall, 96 coastal zip-codes in the state of NJ were used to validate the estimated loss. The model showed a correlation coefficient (CORR = 0.69) between simulated structure losses and NFIP payouts (Fig. 2). In NJ, as of 2019, the total NFIP payout was $3.9 billion USD, in comparison to the estimated total structure loss of $3.6 billion USD (SI Table S3), with an absolute error of 7.7%. This good agreement, plus the good agreement between the simulated and observed surge and wave reported earlier, confirms the validity of our “dynamics-based” loss assessment.
Figure 2

Economic model validation at zip-code resolution. Simulated losses during Sandy in NJ using the USACE damage functions versus FEMA NFIP payouts. Results were aggregated by zip-code and the corresponding correlation coefficient is 0.69 (R2 = 0.47). Validation use transformed structure loss ((PL_{T})) instead of the structure loss ((PL)). The figure is produced using MATLAB R2020 (https://www.mathworks.com/).

Full size image

We define the structural loss reduction ((SLR)) as the structural loss without wetland minus the structural loss with wetland, and the relative structural loss reduction ((RSLR)) as the ratio between (SLR) and the loss with wetland. A state-level analysis of structure loss in NJ showed a (RSLR) of 8.5%, 26.0%, and 52.3% for Sandy, BS storm, and 1% chance flood/wave, respectively (Table S3). Analysis of losses due to flood and wave indicated that for Sandy and the 1% event, most of the loss came from flood, while most of the loss in the BS storm came from waves. Avoided wave-induced loss was comparable to the avoided flood-induced loss during the BS storm and the 1% event, but much higher during Sandy, suggesting that NJ wetlands are more effective in reducing wave-induced loss vs. flood-induced loss. Results from the zip-code scale analysis during Sandy (Fig. 3) showed less variability: for most north NJ, (RSLR) ranged from 10 to 100% except for those along the Hudson River that had small increased loss (negative avoided loss). Most zip-codes in SNJ had (RSLR) between 5% and more than 100%. (RSLR) during the BS storm (Fig. S11) was more notable: ~ 50–100% for north NJ, 0–100% for central NJ. The 1% flood event (Fig. 4) showed an average RSLR  > 25% for NJ.
Figure 3

Percent structural loss reduction during Sandy. Map showing zip-code resolution avoided loss (difference in loss without wetlands and loss with wetlands) during Sandy, as a percentage of the loss of the with-wetland scenario. Dark red values show zip-code with the highest wetland benefit, while dark blue areas have the least benefit. Negative values indicate that the presence of wetland would increase structural losses and positive values indicate that wetland would lower the structural losses. The results are shown for the NJ coastal zip-codes affected by Sandy. This study shows that the percent avoided loss in this figure does not always represent the actual wetland value for loss reduction because areas with few structures and losses could give misleadingly high values of percent avoided loss, as shown in south NJ. The primary purpose is for comparison with a similar figure in NAR17. The map is produced using ESRI ArcGIS Pro 2.7 (https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview).

Full size image

Figure 4

Effect of wetlands on structural losses over zip code scale during the 1% annual event. Map showing zip code resolution difference in losses if the wetlands were absent, as a percentage of the wetland present scenario. Dark red values show zip code with the highest benefit of having wetlands while dark blue areas show the least benefited area. Negative values indicate that the presence of wetland would increase structural losses and positive values indicate that wetland would lower the structural losses. The map is produced using ESRI ArcGIS Pro 2.7 (https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview).

Full size image

The above results showed that the value of coastal wetlands for flood/wave protection varies significantly with the storm. While wetlands may be more effective in reducing wave loss in some storms but flood loss in other storms, they may be ineffective in extreme storms. The 1% annual chance flood and wave event, which resulted from an ensemble of many less extreme but more frequent storms28, provides a more reasonable integrated scenario for the loss analysis. This is similar to the preferred use of the 1% flood map, instead of the flood map associated with a single design storm, for assessing the flood risk in any coastal region.
As shown in Fig. 5, in zip-codes with larger wetland coverage area in SNJ, wetlands could only prevent  $60,000 annual loss. On the other hand, wetlands in north NJ zip-codes with smaller wetland coverage area would increase the loss. The annual loss in most of the zip-codes is  More

A new biodiversity decade is about to start and hopefully will achieve just progress for both people and the planet.

In October 2010, the Convention on Biological Diversity (CBD) signed up to a promising 10-year framework, the Strategic Plan for Biodiversity 2011–2020, which represented a major step to ensure that by 2050 biodiversity would be ‘valued, conserved, restored and wisely used’ in a world where people live in harmony with nature, according to the vision stated in the strategic plan. Underpinning the plan were the so-called 20 Aichi targets aiming to articulate the specific areas of intervention to achieve a set of broad strategic goals — addressing the causes of biodiversity loss and promoting its sustainable use, safeguarding ecosystems, ensuring all would benefit from biodiversity and promoting implementation through participatory planning. The CBD referred to the targets as ambitious but achievable. A decade later, in October 2020, signatories to the CBD were due to discuss progress towards the Aichi targets at the fifteenth Conference of the Parties (COP 15) in Kunming, China, but the meeting was postponed until May 2021 due to the impacts of the COVID-19 pandemic. In the lead up to the meeting, research efforts and wider debates about the Aichi targets have intensified and have all sent a strong and unequivocal message: the state of biodiversity is now worrying more than ever; the Aichi targets have largely not been met. What went wrong? Some feel the problem lays with the ever-lasting challenge of measuring living nature, others refer to inadequate national implementation strategies. Answering the question for CBD members is of critical importance. As COP 15 to the CBD is also the gathering where the parties will agree on a post-2020 biodiversity framework, it is crucial for them to learn from past mistakes. Indeed, the parties have been working to define a new framework and a new set of targets, this time for 2030. Back in October, in preparation for the COP, a revised draft of the new set of biodiversity targets was released. Waiting for the final deliberations in China, some experts have already raised concerns about the effectiveness of the proposed new framework. More specifically, some point to the lack of recognition of the role of agro-ecosystems in conservation, reflecting on the challenges of protecting more land in regions under pressure to feed a growing population (for example, Sub-Saharan Africa). Aichi target 11 required that by 2020 at least 17% of terrestrial areas and 10% of marine areas across the world, especially those of critical importance for biodiversity, should be conserved and equitably managed. Alarmed by lack of progress towards the target, prominent conservationists proposed to ramp up the ambition and protect half of the Earth’s surface to halt biodiversity decline. A study by Mehrabi and colleagues assessed the food production costs of the proposal and found that globally about 23–25% of non-food calories and 3–29% of food calories could be lost. Another study by Schleicher and co-authors found that in the event of including all ecoregions in the proposal there would be substantial socio-economic impacts as over one billion people currently live in areas that would be protected under Half Earth. Both articles, part of a joint collection with Nature Ecology & Evolution, shed light on the challenges of reconciling development needs with biodiversity protection. Some possible solutions to this challenge are controversial; for example, biodiversity offsetting, which refers to actions intended to deliver biodiversity gains as a way of compensating for development impacts. In another study in the collection, Damiens et al. analysed policy documents produced between 1958 and 2019 and showed that biodiversity offsetting was historically promoted through approaches encouraging economic growth with no consideration of biocultural limits. They showed also that, recently, more transformative approaches include offsetting as a tool to transition towards economic systems respectful of planetary boundaries, but their success hinges on realizing quite challenging governance changes.

Credit: Ines Porada / Alamy Stock Photo

Realizing the CBD vision by 2050 requires acknowledging the need of transformative change. What does it mean in practice? It means understanding as best we can the complexity of human–nature interactions. But it also means accepting that there is a lot we don’t know and most probably won’t have enough time to know if we don’t act now. Against this backdrop, the biodiversity community can influence the debate about what should happen in order to sustainably manage living nature. How? Perhaps, as Wyborn and colleagues suggest in their paper, also part of the collection, they could use some imagination to visualize how the future will unfold from the choices made today. There are multiple possible paths to set the world on course to live in harmony with nature — none is perfect, all are possible. This variety reflects diverse values and relations with biodiversity that in turn drive choices. Rather than shying away from it, the biodiversity community should embrace the challenge fully in order to define new priorities for both research and action. More

1.
Slobodkin LB. Prudent predation does not require group selection. Am Nat. 1974;108:665–78.
Article  Google Scholar 
2.
Williams PD. Unhealthy herds: some epidemiological consequences of host heterogeneity in predator-host-parasite systems. J Theor Biol. 2008;253:500–7.
Article  Google Scholar 

3.
Packer C, Holt RD, Hudson PJ, Lafferty KD, Dobson AP. Keeping the herds healthy and alert: Implications of predator control for infectious disease. Ecol Lett. 2003;6:797–802.
Article  Google Scholar 

4.
Lima-Mendez G, Faust K, Henry N, Decelle J, Colin S, Carcillo F, et al. Determinants of community structure in the global plankton interactome. Science 2015;348:1262073.
Article  Google Scholar 

5.
Skovgaard A. Dirty tricks in the plankton: Diversity and role of marine parasitic protists. Acta Protozool. 2014;53:51–62.
Google Scholar 

6.
Jephcott TG, Sime-Ngando T, Gleason FH, Macarthur DJ. Host-parasite interactions in food webs: Diversity, stability, and coevolution. Food Webs. 2016;6:1–8.
Article  Google Scholar 

7.
Nelson DM, Tréguer P, Brzezinski MA, Leynaert A, Quéguiner B. Production and dissolution of biogenic silica in the ocean: Revised global estimates, comparison with regional data and relationship to biogenic sedimentation. Glob Biogeochem Cycles. 1995;9:359–72.
CAS  Article  Google Scholar 

8.
Timmermans KR, Veldhuis MJ, Brussaard CP. Cell death in three marine diatom species in response to different irradiance levels, silicate, or iron concentrations. Aquat Micro Ecol. 2007;46:253–61.
Article  Google Scholar 

9.
Pinto E, Van Nieuwerburgh L, De Barros MP, Pedersén M, Colepicolo P, Snoeijs P. Density-dependent patterns of thiamine and pigment production in the diatom Nitzschia microcephala. Phytochemistry. 2003;63:155–63.
CAS  Article  Google Scholar 

10.
Manoylov KM. Intra- and interspecific competition for nutrients and light in diatom cultures. J Freshw Ecol. 2009;24:145–57.
Article  Google Scholar 

11.
Houston DC, Cooper JE. The digestive tract of the whiteback griffon vulture and its role in disease transmission among wild ungulates. J Wildl Dis. 1975;11:306–13.
CAS  Article  Google Scholar 

12.
Schaller G. The Serengeti lion: a study of predator-prey relations. London: University of Chicago Press; 1972.

13.
Krumm CE, Conner MM, Hobbs NT, Hunter DO, Miller MW. Mountain lions prey selectively on prion-infected mule deer. Biol Lett. 2010;6:209–11.
Article  Google Scholar 

14.
Pole A, Gordon IJ, Gorman ML, MacAskill M. Prey selection by African wild dogs (Lycaon pictus) in southern Zimbabwe. J Zool. 2004;262:207–15.
Article  Google Scholar 

15.
Husseman JS, Murray DL, Power G, Mack C, Wenger CR, Quigley H. Assessing differential prey selection patterns between two sympatric large carnivores. Oikos. 2003;101:591–601.
Article  Google Scholar 

16.
Lafferty KD. Fishing for lobsters indirectly increases epidemics in sea urchins. Ecol Appl. 2004;14:1566–73.
Article  Google Scholar 

17.
Duffy MA, Hall SR, Tessier AJ, Huebner M. Selective predators and their parasitized prey: Are epidemics in zooplankton under top-down control? Limnol Oceanogr. 2005;50:412–20.
Article  Google Scholar 

18.
Hudson PJ, Dobson AP, Newborn D. Do parasites make prey vulnerable to predation? Red grouse and parasites. J Anim Ecol. 1992;61:681.
Article  Google Scholar  More