Ecology

1.Golden, C. D. et al. Nutrition: Fall in fish catch threatens human health. Nature 534, 317–320 (2016).PubMed 
Article 
PubMed Central 

Google Scholar 
2.Halpern, B. S. et al. Spatial and temporal changes in cumulative human impacts on the world’s ocean. Nat. Commun. 6, 7615 (2015).3.The Ocean Economy in 2030 (OECD, 2016).4.Boonstra, W. J., Valman, M. & Björkvik, E. A sea of many colours – how relevant is blue growth for capture fisheries in the global north, and vice versa? Mar. Policy 87, 340–349 (2018).Article 

Google Scholar 
5.Bennett, N. J. et al. Towards a sustainable and equitable blue economy. Nat. Sustain. 2, 991–993 (2019).Article 

Google Scholar 
6.Nash, K. L. et al. Planetary boundaries for a blue planet. Nat. Ecol. Evol. 1, 1625–1634 (2017).PubMed 
Article 
PubMed Central 

Google Scholar 
7.Maxwell, S. M. et al. Cumulative human impacts on marine predators. Nat. Commun. 4, 2688 (2013).8.Zero Draft of the Post-2020 Global Biodiversity Framework (CBD, 2020).9.Update of the Zero Draft of the Post-2020 Global Biodiversity Framework (CBD, 2020).10.Mace, G. M. et al. Aiming higher to bend the curve of biodiversity loss. Nat. Sustain. 1, 448–451 (2018).Article 

Google Scholar 
11.Leclere, D. et al. Bending the curve of terrestrial biodiversity needs an integrated strategy. Nature 585, 551–556 (2020).CAS 
PubMed 
PubMed Central 

Google Scholar 
12.Bull, J. W. et al. Net positive outcomes for nature. Nat. Ecol. Evol. 4, 4–7 (2020).PubMed 
Article 
PubMed Central 

Google Scholar 
13.Milner-Gulland, E. J. et al. Four steps for the Earth: mainstreaming the post-2020 global biodiversity framework. One Earth 4, 75–87 (2020).Article 

Google Scholar 
14.Davies, R. W. D., Cripps, S. J. J., Nickson, A. & Porter, G. Defining and estimating global marine fisheries bycatch. Mar. Policy 33, 661–672 (2009).Article 

Google Scholar 
15.Hall, M. A., Alverson, D. L. & Metuzals, K. I. By-catch: problems and solutions. Mar. Pollut. Bull. 41, 204–219 (2000).CAS 
Article 

Google Scholar 
16.Crowder, L. B. & Murawski, S. A. Fisheries bycatch: implications for management. Fisheries 23, 8–17 (1998).Article 

Google Scholar 
17.Branch, T. A., Rutherford, K. & Hilborn, R. Replacing trip limits with individual transferable quotas: implications for discarding. Mar. Policy 30, 281–292 (2006).Article 

Google Scholar 
18.Lewison, R. L., Crowder, L. B., Read, A. J. & Freeman, S. A. Understanding impacts of fisheries bycatch on marine megafauna. Trends Ecol. Evol. 19, 598–604 (2004).Article 

Google Scholar 
19.Gilman, E. et al. Reducing sea turtle interactions in the Hawaii-based longline swordfish fishery. Biol. Conserv. 139, 19–28 (2007).Article 

Google Scholar 
20.Watson, J. T., Essington, T. E., Lennert-Cody, C. E. & Hall, M. A. Trade-offs in the design of fishery closures: management of silky shark bycatch in the eastern Pacific Ocean tuna fishery. Conserv. Biol. 23, 626–635 (2009).PubMed 
Article 
PubMed Central 

Google Scholar 
21.Campbell, L. M. & Cornwell, M. L. Human dimensions of bycatch reduction technology: current assumptions and directions for future research. Endang. Species Res. 5, 325–334 (2008).Article 

Google Scholar 
22.Hall, M. A. On bycatches. Rev. Fish Biol. Fish. 6, 319–352 (1996).Article 

Google Scholar 
23.Smith, V. L. On models of commercial fishing. J. Polit. Econ. 77, 181–198 (1969).Article 

Google Scholar 
24.Innes, J., Pascoe, S., Wilcox, C., Jennings, S. & Paredes, S. Mitigating undesirable impacts in the marine environment: a review of market-based management measures. Front. Mar. Sci. 2, 76 (2015).Article 

Google Scholar 
25.Nyborg, K. et al. Social norms as solutions. Science 354, 42–43 (2016).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
26.Bowles, S. & Polanía-Reyes, S. Economic incentives and social preferences: substitutes or complements? J. Econ. Lit. 50, 368–425 (2012).Article 

Google Scholar 
27.Gneezy, U., Meier, S. & Rey-Biel, P. When and why incentives (don’t) work to modify behavior. J. Econ. Perspect. 25, 191–210 (2011).Article 

Google Scholar 
28.Fulton, E. A., Smith, A. D. M., Smith, D. C. & Van Putten, I. E. Human behaviour: the key source of uncertainty in fisheries management. Fish Fish. 12, 2–17 (2011).Article 

Google Scholar 
29.Sumaila, U. R., Lam, V., Le Manach, F., Swartz, W. & Pauly, D. Global fisheries subsidies: an updated estimate. Mar. Policy 69, 189–193 (2016).Article 

Google Scholar 
30.Bladon, A. J., Short, K. M., Mohammed, E. Y. & Milner-Gulland, E. J. Payments for ecosystem services in developing world fisheries. Fish Fish. 17, 839–859 (2016).Article 

Google Scholar 
31.Deutza, A. et al. Financing Nature: Closing the Global Biodiversity Financing Gap (The Paulson Institute, The Nature Conservancy, and the Cornell Atkinson Center for Sustainability, 2020).32.Dutton, P. H. & Squires, D. Reconciling biodiversity with fishing: a holistic strategy for Pacific sea turtle recovery. Ocean Dev. Int. Law 39, 200–222 (2008).Article 

Google Scholar 
33.Pascoe, S. et al. Use of incentive-based management systems to limit bycatch and discarding. Int. Rev. Environ. Resour. Econ. 4, 123–161 (2010).Article 

Google Scholar 
34.Wilcox, C. & Donlan, C. J. Compensatory mitigation as a solution to fisheries bycatch–biodiversity conservation conflicts. Front. Ecol. Environ. 5, 325–331 (2007).Article 

Google Scholar 
35.Dutton, P. H. & Squires, D. in Conservation of Pacific Sea Turtles (eds Dutton, P. H. et al.) 37–59 (Univ. Hawaii Press, 2011).36.Sumaila, U. R. et al. Ocean Finance: Financing the Transition to a Sustainable Ocean Economy (World Resources Institute, 2020).37.Squires, D., Restrepo, V., Garcia, S. & Dutton, P. Fisheries bycatch reduction within the least-cost biodiversity mitigation hierarchy: conservatory offsets with an application to sea turtles. Mar. Policy 93, 55–61 (2018).Article 

Google Scholar 
38.Finkelstein, M. et al. Evaluating the potential effectiveness of compensatory mitigation strategies for marine bycatch. PLoS ONE 3, e2480 (2008).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
39.Žydelis, R., Wallace, B. P., Gilman, E. L. & Werner, T. B. Conservation of marine megafauna through minimization of fisheries bycatch. Conserv. Biol. 23, 608–616 (2009).PubMed 
Article 
PubMed Central 

Google Scholar 
40.Blomquist, J., Bartolino, V. & Waldo, S. Price premiums for providing eco-labelled seafood: evidence from MSC-certified cod in Sweden. J. Agric. Econ. 66, 690–704 (2015).Article 

Google Scholar 
41.Roheim, C. A., Bush, S. R., Asche, F., Sanchirico, J. N. & Uchida, H. Evolution and future of the sustainable seafood market. Nat. Sustain. 1, 392–398 (2018).Article 

Google Scholar 
42.Shumway, N., Watson, J. E. M., Saunders, M. I. & Maron, M. The risks and opportunities of translating terrestrial biodiversity offsets to the marine realm. BioScience 68, 125–133 (2018).Article 

Google Scholar 
43.Bull, J. W. & Strange, N. The global extent of biodiversity offset implementation under no net loss policies. Nat. Sustain. 1, 790–798 (2018).Article 

Google Scholar 
44.Gjertsen, H., Squires, D., Dutton, P. H. & Eguchi, T. Cost-effectiveness of alternative conservation strategies with application to the Pacific leatherback turtle. Conserv. Biol. 28, 140–149 (2014).PubMed 
Article 
PubMed Central 

Google Scholar 
45.Segerson, K. in Conservation of Pacific Sea Turtles (eds Dutton, P. H. et al.) 370–395 (Univ. Hawaii Press, 2011).46.Bull, J. W., Suttle, K. B., Gordon, A., Singh, N. J. & Milner-Gulland, E. J. Biodiversity offsets in theory and practice. Oryx 47, 369–380 (2013).Article 

Google Scholar 
47.zu Ermgassen, S. O. S. E. et al. The ecological outcomes of biodiversity offsets under ‘no net loss’ policies: a global review. Conserv. Lett. 12, e12664 (2019).48.zu Ermgassen, S. O. S. E. et al. Exploring the ecological outcomes of mandatory biodiversity net gain using evidence from early-adopter jurisdictions in England. Preprint at SocArXiv https://doi.org/10.31235/osf.io/tw6nr (2021).49.Segerson, K. Voluntary approaches to environmental protection and resource management. Annu. Rev. Resour. Econ. 5, 161–180 (2013).Article 

Google Scholar 
50.Kotchen, M. J. Voluntary- and information-based approaches to environmental management: a public economics perspective. Rev. Environ. Econ. Policy 7, 276–295 (2013).Article 

Google Scholar 
51.Janisse, C., Squires, D., Seminoff, J. A. & Dutton, P. H. in Handbook of Marine Fisheries Conservation and Management, (eds Grafton, Q. et al.) 231–240 (Oxford Univ. Press, 2010).52.Squires, D. & Garcia, S. The least-cost biodiversity impact mitigation hierarchy with a focus on marine fisheries and bycatch issues. Conserv. Biol. 32, 989–997 (2018).PubMed 
Article 
PubMed Central 

Google Scholar 
53.Jacob, C. et al. Marine biodiversity offsets: pragmatic approaches toward better conservation outcomes. Conserv Lett. 13, 1–12 (2020).Article 

Google Scholar 
54.Lent, R. & Squires, D. Reducing marine mammal bycatch in global fisheries: an economics approach. Deep Sea Res. Pt II 140, 268–277 (2017).Article 

Google Scholar 
55.Zhou, R. & Segerson, K. Individual vs. collective approaches to fisheries management. Mar. Resour. Econ. 31, 165–192 (2016).Article 

Google Scholar 
56.Yagi, N., Clark, M. L., Anderson, L. G., Arnason, R. & Metzner, R. Applicability of individual transferable quotas (ITQs) in Japanese fisheries: a comparison of rights-based fisheries management in Iceland, Japan, and United States. Mar. Policy 36, 241–245 (2012).Article 

Google Scholar 
57.Kotchen, M. J. & Segerson, K. On the use of group performance and rights for environmental protection and resource management. Proc. Natl Acad. Sci. USA 116, 5285–5292 (2019).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
58.Stewart, A. J. & Plotkin, J. B. Collapse of cooperation in evolving games. Proc. Natl Acad. Sci. USA 111, 17558–17563 (2014).59.Mani, A., Rahwan, I. & Pentland, A. Inducing peer pressure to promote cooperation. Sci. Rep. 3, 1735 (2013).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
60.Kroeger, T. The quest for the ‘optimal’ payment for environmental services program: ambition meets reality, with useful lessons. Policy Econ. 37, 65–74 (2013).Article 

Google Scholar 
61.Ledoux, L. & Turner, R. K. Valuing ocean and coastal resources: a review of practical examples and issues for further action. Ocean Coast. Manag. 45, 583–616 (2002).Article 

Google Scholar 
62.Booth, H., Squires, D. & Milner-Gulland, E. J. The mitigation hierarchy for sharks: a risk-based framework for reconciling trade-offs between shark conservation and fisheries objectives. Fish Fish. 21, 269–289 (2019).Article 

Google Scholar 
63.Bull, J. W. & Milner-Gulland, E. Choosing prevention or cure when mitigating biodiversity loss: trade-offs under ‘no net loss’ policies. J. Appl. Ecol. 57, 354–366 (2020).Article 

Google Scholar 
64.Norton, D. A. & Warburton, B. The potential for biodiversity offsetting to fund effective invasive species control. Conserv. Biol. 29, 5–11 (2015).PubMed 
Article 
PubMed Central 

Google Scholar 
65.Holmes, N. D. et al. The potential for biodiversity offsetting to fund invasive species eradications on islands. Conserv. Biol. 30, 425–427 (2016).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
66.Gallagher, A. J. & Hammerschlag, N. Global shark currency: the distribution frequency and economic value of shark ecotourism. Curr. Issues Tour. 14, 797–812 (2011).Article 

Google Scholar 
67.Mustika, P. L. K., Ichsan, M. & Booth, H. The economic value of shark and ray tourism in Indonesia and its role in delivering conservation outcomes. Front. Mar. Sci. 7, 261 (2020).Article 

Google Scholar 
68.O’Malley, M. P., Lee-Brooks, K. & Medd, H. B. The global economic impact of manta ray watching tourism. PLoS ONE 8, e65051 (2013).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
69.Vianna, G. M. S. et al. Shark-diving tourism as a financing mechanism for shark conservation strategies in Malaysia. Mar. Policy 94, 220–226 (2018).Article 

Google Scholar 
70.Swimmer, Y. et al. Sea turtle bycatch mitigation in U.S. longline fisheries. Front. Mar. Sci. 4, 260 (2017).Article 

Google Scholar 
71.Croll, D. A. et al. Vulnerabilities and fisheries impacts: the uncertain future of manta and devil rays. Aquat. Conserv. Mar. Freshw. Ecosyst. 26, 562–575 (2016).Article 

Google Scholar 
72.Anderson, R. C., Adam, M. S., Kitchen-Wheeler, A. M. & Stevens, G. Extent and economic value of manta ray watching in Maldives. Tour. Mar. Environ. 7, 15–27 (2011).Article 

Google Scholar 
73.Farley, J. Ecosystem services: the economics debate. Ecosyst. Serv. 1, 40–49 (2012).Article 

Google Scholar 
74.Kenter, J. O., Hyde, T., Christie, M. & Fazey, I. The importance of deliberation in valuing ecosystem services in developing countries—evidence from the Solomon Islands. Glob. Environ. Change 21, 505–521 (2011).Article 

Google Scholar 
75.Gilman, E. et al. Robbing Peter to pay Paul: replacing unintended cross-taxa conflicts with intentional tradeoffs by moving from piecemeal to integrated fisheries bycatch management. Rev. Fish Biol. Fish. 29, 93–123 (2019).Article 

Google Scholar 
76.Dissou, Y. & Siddiqui, M. S. Can carbon taxes be progressive?. Energy Econ. 42, 88–100 (2014).Article 

Google Scholar 
77.Engel, S., Pagiola, S. & Wunder, S. Designing payments for environmental services in theory and practice: an overview of the issues. Ecol. Econ. 65, 663–674 (2008).Article 

Google Scholar 
78.Bene, C. Small-Scale Fisheries: Assessing Their Contribution to Rural Livelihoods in Developing Countries FAO Fisheries Circular No. 1008 (FAO, 2006).79.Arlidge, W. N. S. et al. A mitigation hierarchy approach for managing sea turtle captures in small-scale fisheries. Front. Mar. Sci. 7, 49 (2020).Article 

Google Scholar 
80.Levi, M., Sacks, A. & Tyler, T. Conceptualizing legitimacy, measuring legitimating beliefs. Am. Behav. Sci. 53, 354–375 (2009).Article 

Google Scholar 
81.Oyanedel, R., Gelcich, S. & Milner-Gulland, E. J. Motivations for (non-)compliance with conservation rules by small-scale resource users. Conserv. Lett. 15, e12725 (2020).
Google Scholar 
82.Pakiding, F. et al. Community engagement: an integral component of a multifaceted conservation approach for the transboundary western Pacific leatherback. Front. Mar. Sci. 7, 756 (2020).Article 

Google Scholar 
83.Long-Term Strategic Directions to the 2050 Vision for Biodiversity, Approaches to Living in Harmony with Nature and Preparation for the Post-2020 Global Biodiversity Framework Report No. CBD/COP/14/9 (CBD, 2018).84.Berkes, F. et al. Globalization, roving bandits, and marine resources. Science 311, 1557–1558 (2006).CAS 
Article 

Google Scholar 
85.Ban, N. C., Adams, V., Pressey, R. L. & Hicks, J. Promise and problems for estimating management costs of marine protected areas. Conserv. Lett. 4, 241–252 (2011).Article 

Google Scholar 
86.Arias, A., Pressey, R. L., Jones, R. E., Álvarez-Romero, J. G. & Cinner, J. E. Optimizing enforcement and compliance in offshore marine protected areas: a case study from Cocos Island, Costa Rica. Oryx 50, 18–26 (2016).Article 

Google Scholar 
87.Bartholomew, D. C. et al. Remote electronic monitoring as a potential alternative to on-board observers in small-scale fisheries. Biol. Conserv. 219, 35–45 (2018).Article 

Google Scholar 
88.Mangi, S. C., Dolder, P. J., Catchpole, T. L., Rodmell, D. & de Rozarieux, N. Approaches to fully documented fisheries: practical issues and stakeholder perceptions. Fish Fish. 16, 426–452 (2015).Article 

Google Scholar 
89.Harper, L. R. et al. Prospects and challenges of environmental DNA (eDNA) monitoring in freshwater ponds. Hydrobiologia 826, 25–41 (2019).90.Russo, T. et al. All is fish that comes to the net: metabarcoding for rapid fisheries catch assessment. Ecol. Appl. 31, e02273 (2021).PubMed 
Article 
PubMed Central 

Google Scholar 
91.Cardeñosa, D., Gollock, M. J. & Chapman, D. D. Development and application of a novel real‐time polymerase chain reaction assay to detect illegal trade of the European eel (Anguilla anguilla). Conserv. Sci. Prac. 1, e39 (2019).Article 

Google Scholar 
92.Griffiths, V. F., Bull, J. W., Baker, J. & Milner-Gulland, E. J. No net loss for people and biodiversity. Conserv. Biol. 33, 76–87 (2019).PubMed 
Article 
PubMed Central 

Google Scholar 
93.Milner-Gulland, E. J. et al. Translating the terrestrial mitigation hierarchy to marine megafauna by-catch. Fish Fish. 19, 547–561 (2018).Article 

Google Scholar 
94.Republic of Namibia Ministry of Fisheries and Marine Resources Annual Report 2012–2013 (MFMR, 2013).95.Sanchirico, J. N., Holland, D., Quigley, K. & Fina, M. Catch-quota balancing in multispecies individual fishing quotas. Mar. Policy 30, 767–785 (2006).Article 

Google Scholar 
96.Walker, S. & Townsend, R. Economic analysis of New Zealand’s deemed value system. In Proc. of the Fourteenth Biennial Conference of the International Institute of Fisheries Economics & Trade, July 22–25, 2008, Nha Trang, Vietnam: Achieving a Sustainable Future: Managing Aquaculture, Fishing, Trade and Development 1–11 (Oregon State Univ., 2008).97.Sanchirico, J. N. Managing marine capture fisheries with incentive based price instruments. Public Finance Manag. 3, 67–93 (2003).
Google Scholar 
98.Pascoe, S., Wilcox, C. & Donlan, C. J. Biodiversity offsets: a cost-effective interim solution to seabird bycatch in fisheries? PLoS ONE 6, e25762 (2011).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
99.Mukherjee, Z. Controlling stochastic externalities with penalty threats: the case of bycatch. Environ. Econ. Policy Stud. 18, 93–113 (2016).Article 

Google Scholar 
100.Androkovich, R. A. & Stollery, K. R. A stochastic dynamic programming model of bycatch control in fisheries. Mar. Resour. Econ. 9, 19–30 (1994).Article 

Google Scholar 
101.Herrera, G. E. Stochastic bycatch, informational asymmetry, and discarding. J. Environ. Econ. Manag. 49, 463–483 (2005).Article 

Google Scholar 
102.Singh, R. & Weninger, Q. Bioeconomies of scope and the discard problem in multiple-species fisheries. J. Environ. Econ. Manag. 58, 72–92 (2009).Article 

Google Scholar 
103.Pascoe, S., Cannard, T. & Steven, A. Offset payments can reduce environmental impacts of urban development. Environ. Sci. Policy 100, 205–210 (2019).Article 

Google Scholar 
104.Schouten, G. & Glasbergen, P. Creating legitimacy in global private governance: the case of the roundtable on sustainable palm oil. Ecol. Econ. 70, 1891–1899 (2011).Article 

Google Scholar 
105.Ruysschaert, D. & Salles, D. in The Anthropology of Conservation NGOs (eds Larsen, P. B. & Brockington, D.) 121–149 (Palgrave Macmillan, 2018).106.Right Whales and Entanglements: More on How NOAA Makes Decisions (NOAA Fisheries, 2019); https://www.fisheries.noaa.gov/new-england-mid-atlantic/marine-mammal-protection/right-whales-and-entanglements-more-how-noaa#right-whales-and-the-lobster-fishery107.Jouffray, J. B., Crona, B., Wassénius, E., Bebbington, J. & Scholtens, B. Leverage points in the financial sector for seafood sustainability. Sci. Adv. 5, eaax3324 (2019).PubMed 
PubMed Central 
Article 

Google Scholar 
108.Österblom, H. et al. Emergence of a global science-business initiative for ocean stewardship. Proc. Natl Acad. Sci. USA 114, 9038–9043 (2017).PubMed 
Article 
CAS 
PubMed Central 

Google Scholar 
109.Deforestation-Free Supply Chains: From Commitments to Action (CDP, 2014).110.Donofrio, S., Rothrock, P. & Leonard, J. Supply Change: Tracking Corporate Commitments to Deforestation-Free Supply Chains (Forest Trends, 2017).111.Österblom, H. et al. Transnational corporations as ‘keystone actors’ in marine ecosystems. PLoS ONE 10, e0127533 (2015).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
112.Galland, G., Rogers, A. & Nickson, A. Netting Billions: A Global Valuation of Tuna (The Pew Charitible Trusts, 2016).113.Zeller, D., Cashion, T., Palomares, M. & Pauly, D. Global marine fisheries discards: a synthesis of reconstructed data. Fish Fish. 19, 30–39 (2018).Article 

Google Scholar 
114.Balmford, A., Gravestock, P., Hockley, N., McClean, C. J. & Roberts, C. M. The worldwide costs of marine protected areas. Proc. Natl Acad. Sci. USA 101, 9694–9697 (2004).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
115.Engel, S. & Palmer, C. Payments for environmental services as an alternative to logging under weak property rights: the case of Indonesia. Ecol. Econ. 65, 799–809 (2008).Article 

Google Scholar 
116.Bonham, C. et al. Conservation trust funds, protected area management effectiveness and conservation outcomes: lessons from the global conservation fund. Parks 20, 89–100 (2014).Article 

Google Scholar 
117.Angelsen, A. et al. (eds) Realising REDD+: National Strategy and Policy Options (CIFOR, 2009).118.Spergel, B. & Mikitin, K. Practice Standards for Conservation Trust Funds (CFA, 2013).119.Progress Summary of 2014–15 ISSF Funded Marine Turtle Projects (ISSF, 2016). More

1.Schaal, B. Plants and people: our shared history and future. Plants People Planet 1, 14–19 (2019).Article 

Google Scholar 
2.Bates, D. M. People, plants and genes: the story of crops and humanity. Q. Rev. Biol. 84, 206–207 (2009).3.Nedelcheva, A., Dogan, Y., Obratov-Petkovic, D. & Padure, I. M. The traditional use of plants for handicrafts in southeastern Europe. Hum. Ecol. Interdiscip. J. 39, 813–828 (2011).Article 

Google Scholar 
4.Willes, M. A Shakespearean Botanical (Bodleian Library, 2015).5.Shoemaker, C. A. Plants and human culture. J. Home Consum. Hortic. 1, 3–7 (1994).Article 

Google Scholar 
6.Alfred, J. & Baldwin, I. T. The natural history of model organisms: new opportunities at the wild frontier. eLife 4, e06956 (2015).PubMed Central 
Article 

Google Scholar 
7.Hedges, S. B. The origin and evolution of model organisms. Nat. Rev. Genet. 3, 838–849 (2002).CAS 
PubMed 
Article 

Google Scholar 
8.Clark, J. A. Taxonomic bias in conservation research. Science 297, 191–192 (2002).CAS 
PubMed 
Article 

Google Scholar 
9.Mammola, S. et al. Towards a taxonomically unbiased European Union biodiversity strategy for 2030. Proc. R. Soc. B 287, 20202166 (2020).PubMed 
Article 

Google Scholar 
10.Christenhusz, M. J. M. & Byng, J. W. The number of known plants species in the world and its annual increase. Phytotaxa 261, 201–217 (2016).Article 

Google Scholar 
11.Quijas, S., Schmid, B. & Balvanera, P. Plant diversity enhances provision of ecosystem services: a new synthesis. Basic Appl. Ecol. 11, 582–593 (2010).Article 

Google Scholar 
12.Zak, D. R., Holmes, W. E., White, D. C., Peacock, A. D. & Tilman, D. Plant diversity, soil microbial communities, and ecosystem function: are there any links? Ecology 84, 2042–2050 (2003).Article 

Google Scholar 
13.Barnosky, A. D. et al. Has the Earth’s sixth mass extinction already arrived? Nature 471, 51–57 (2011).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
14.Cardinale, B. J. et al. Biodiversity loss and its impact on humanity. Nature 486, 59–67 (2012).CAS 
Article 

Google Scholar 
15.Ripple, W. J. et al. World scientists’ warning to humanity: a second notice. Bioscience 67, 1026–1028 (2017).Article 

Google Scholar 
16.Balding, M. & Williams, K. J. H. Plant blindness and the implications for plant conservation. Conserv. Biol. 30, 1192–1199 (2016).PubMed 
Article 
PubMed Central 

Google Scholar 
17.Fukushima, C. S., Mammola, S. & Cardoso, P. Global wildlife trade permeates the Tree of Life. Biol. Conserv. 247, 108503 (2020).PubMed 
PubMed Central 
Article 

Google Scholar 
18.Wandersee, J. H. & Schussler, E. E. Preventing plant blindness. Am. Biol. Teach. 61, 82–86 (1999).Article 

Google Scholar 
19.Parsley, K. M. Plant awareness disparity: a case for renaming plant blindness. Plants People Planet 2, 598–601 (2020).Article 

Google Scholar 
20.Villemant, C. et al. The Mercantour/Alpi Marittime All Taxa Biodiversity Inventory (ATBI): achievements and prospects. Zoosystema 37, 667–679 (2015).Article 

Google Scholar 
21.Borcard, D., Legendre, P. & Drapeau, P. Partialling out the spatial component of ecological variation. Ecology 73, 1045–1055 (1992).Article 

Google Scholar 
22.Médail, F. & Verlaque, R. Ecological characteristics and rarity of endemic plants from Southeast France and Corsica: implications for biodiversity conservation. Biol. Conserv. 80, 269–281 (1997).Article 

Google Scholar 
23.Noble, V. & Diadema, K. in La flore des Alpes-Maritimes et de la Principauté de Monaco (eds Noble, V. & Diadema, K.) 57–72 (Naturalia, 2011).24.Zuur, A. F. & Ieno, E. N. A protocol for conducting and presenting results of regression-type analyses. Methods Ecol. Evol. 7, 636–645 (2016).Article 

Google Scholar 
25.Horiguchi, H., Winawer, J., Dougherty, R. F. & Wandell, B. A. Human trichromacy revisited. Proc. Natl Acad. Sci. USA 110, E260–E269 (2013).CAS 
PubMed 
Article 

Google Scholar 
26.Gerl, E. J. & Morris, M. R. The causes and consequences of color vision. Evol. Educ. Outreach 1, 476–486 (2008).Article 

Google Scholar 
27.Bompas, A., Kendall, G. & Sumner, P. Spotting fruit versus picking fruit as the selective advantage of human colour vision. i-Perception 4, 84–94 (2013).PubMed 
PubMed Central 
Article 

Google Scholar 
28.Elliot, A. J. & Maier, M. A. Color psychology: effects of perceiving color on psychological functioning in humans. Annu. Rev. Psychol. 65, 95–120 (2014).PubMed 
Article 

Google Scholar 
29.Chiao, J. Y. et al. Dynamic cultural influences on neural representations of the self. J. Cogn. Neurosci. 22, 1–11 (2010).PubMed 
Article 

Google Scholar 
30.Cotto, O. et al. A dynamic eco-evolutionary model predicts slow response of alpine plants to climate warming. Nat. Commun. 8, 15399 (2017).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
31.Costa, G. C., Nogueira, C., Machado, R. B. & Colli, G. R. Sampling bias and the use of ecological niche modeling in conservation planning: a field evaluation in a biodiversity hotspot. Biodivers. Conserv. 19, 883–899 (2010).Article 

Google Scholar 
32.Tucker, C. M. et al. A guide to phylogenetic metrics for conservation, community ecology and macroecology. Biol. Rev. Camb. Philos. Soc. 92, 698–715 (2017).PubMed 
Article 

Google Scholar 
33.De Boeck, H. J., Liberloo, M., Gielen, B., Nijs, I. & Ceulemans, R. The observer effect in plant science. New Phytol. 177, 579–583 (2008).PubMed 
Article 
PubMed Central 

Google Scholar 
34.Morrison, L. W. Observer error in vegetation surveys: a review. J. Plant Ecol. 9, 367–379 (2016).Article 

Google Scholar 
35.Kéry, M. & Gregg, K. B. Effects of life-state on detectability in a demographic study of the terrestrial orchid Cleistes bifaria. J. Ecol. 91, 265–273 (2003).Article 

Google Scholar 
36.Allioni, C. Flora Pedemontana: Sive Enumeratio Methodica Stirpium Indigenarum Pedemontii Vol. 1 (Joannes Michael Briolus, 1785).37.Aeschimann, D., Rasolofo, N. & Theurillat, J.-P. Analyse de la flore des alpes. 1: Historique et biodiversité. Candollea 66, 27–55 (2011).Article 

Google Scholar 
38.Boakes, E. H. et al. Distorted views of biodiversity: spatial and temporal bias in species occurrence data. PLoS Biol. 8, e1000385 (2010).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
39.Julve, P. Baseflor. Index Botanique, Ecologique et Chorologique de la Flore de France http://philippe.julve.pagesperso-orange.fr/baseflor.xlsx (1998).40.Bartolucci, F. et al. An updated checklist of the vascular flora native to Italy. Plant Biosyst. 152, 179–303 (2018).Article 

Google Scholar 
41.Web of Science (Clarivate Analytics, accessed 16 March 2020); https://www.webofknowledge.com42.Kalwij, J. M. Review of ‘The Plant List, a working list of all plant species’. J. Veg. Sci. 23, 998–1002 (2012).Article 

Google Scholar 
43.Konno, K. et al. Ignoring non‐English‐language studies may bias ecological meta‐analyses. Ecol. Evol. 10, 6373–6384 (2020).44.Heaton, L., Millerand, F. & Proulx, S. Tela Botanica: une fertilisation croisée des amateurs et des experts. Hermès 57, 61–68 (2010).45.Lauber, K., Wagner, G. & Gygax, A. Flora Helvetica: Illustrierte Flora der Schweiz (Haupt Verlag, 2018).46.Landolt, E. et al. Flora Indicativa: Okologische Zeigerwerte und Biologische Kennzeichen zur Flora der Schweiz und der Alpen (Haupt, 2010).47.The IUCN Red List of Threatened Species (IUCN, 2020).48.Global Biodiversity Information Facility (2020); https://www.gbif.org49.Beck, J., Böller, M., Erhardt, A. & Schwanghart, W. Spatial bias in the GBIF database and its effect on modeling species’ geographic distributions. Ecol. Inform. 19, 10–15 (2014).Article 

Google Scholar 
50.Shirey, V., Belitz, M. W., Barve, V. & Guralnick, R. A complete inventory of North American butterfly occurrence data: narrowing data gaps, but increasing bias. Ecography 44, 537–547 (2021).Article 

Google Scholar 
51.R. Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2019).52.Zuur, A. F., Ieno, E. N. & Elphick, C. S. A protocol for data exploration to avoid common statistical problems. Methods Ecol. Evol. 1, 3–14 (2010).Article 

Google Scholar 
53.Brooks, T. M. et al. Measuring terrestrial area of habitat (AOH) and its utility for the IUCN Red List. Trends Ecol. Evol. 34, 977–986 (2019).PubMed 
Article 

Google Scholar 
54.Bartoń, K. MuMIn: multi-model inference. R package version 1.43.17 https://cran.r-project.org/package=MuMIn (2020).55.Barbosa, A. M., Real, R., Munoz, A. R. & Brown, J. A. New measures for assessing model equilibrium and prediction mismatch in species distribution models. Divers. Distrib. 19, 1333–1338 (2013).Article 

Google Scholar 
56.Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1 (2015).Article 

Google Scholar 
57.Blasco-Moreno, A., Pérez-Casany, M., Puig, P., Morante, M. & Castells, E. What does a zero mean? Understanding false, random and structural zeros in ecology. Methods Ecol. Evol. 10, 949–959 (2019).Article 

Google Scholar 
58.Johnson, J. B. & Omland, K. S. Model selection in ecology and evolution. Trends Ecol. Evol. 19, 101–108 (2004).PubMed 
Article 
PubMed Central 

Google Scholar 
59.Lüdecke, D., Ben-Shachar, M., Patil, I., Waggoner, P. & Makowski, D. Assessment, testing and comparison of statistical models using R. Preprint at PsyArXiv https://doi.org/10.31234/osf.io/vtq8f (2021). More

1.Bowman, D. M. J. S. et al. Fire in the Earth system. Science 324, 481–484 (2009).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
2.Davis, K. T. et al. Wildfires and climate change push low-elevation forests across a critical climate threshold for tree regeneration. Proc. Natl Acad. Sci. USA 116, 6193–6198 (2019).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
3.Enright, N. J., Fontaine, J. B., Bowman, D. M. J. S., Bradstock, R. A. & Williams, R. J. Interval squeeze: altered fire regimes and demographic responses interact to threaten woody species persistence as climate changes. Front. Ecol. Environ. 13, 265–272 (2015).Article 

Google Scholar 
4.Abatzoglou, J. T. & Williams, A. P. Impact of anthropogenic climate change on wildfire across western US forests. Proc. Natl Acad. Sci. USA 113, 11770–11775 (2016).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
5.Bowman, D. M. J. S. et al. Vegetation fires in the Anthropocene. Nat. Rev. Earth Environ. 1, 500–515 (2020).Article 

Google Scholar 
6.Keeley, J. E., van Mantgem, P. & Falk, D. A. Fire, climate and changing forests. Nat. Plants 5, 774–775 (2019).PubMed 
Article 
PubMed Central 

Google Scholar 
7.Lindenmayer, D. B., Kooyman, R. M., Taylor, C., Ward, M. & Watson, J. E. Recent Australian wildfires made worse by logging and associated forest management. Nat. Ecol. Evol. 4, 898–900 (2020).PubMed 
Article 
PubMed Central 

Google Scholar 
8.Murphy, B. P. et al. Fire regimes of Australia: a pyrogeographic model system. J. Biogeogr. 40, 1048–1058 (2013).Article 

Google Scholar 
9.Poulos, H. M., Barton, A. M., Slingsby, J. A. & Bowman, D. M. J. S. Do mixed fire regimes shape plant flammability and post-fire recovery strategies? Fire 1, 39 (2018).Article 

Google Scholar 
10.Cawson, J. G. et al. Exploring the key drivers of forest flammability in wet eucalypt forests using expert-derived conceptual models. Landsc. Ecol. 35, 1775–1798 (2020).Article 

Google Scholar 
11.Thomas, P. B., Watson, P. J., Bradstock, R. A., Penman, T. D. & Price, O. Modelling surface fine fuel dynamics across climate gradients in eucalypt forests of south‐eastern Australia. Ecography 37, 827–837 (2014).Article 

Google Scholar 
12.Bennett, L. T. et al. Mortality and recruitment of fire-tolerant eucalypts as influenced by wildfire severity and recent prescribed fire. For. Ecol. Manag. 380, 107–117 (2016).Article 

Google Scholar 
13.Fairman, T. A., Bennett, L. T., Tupper, S. & Nitschke, C. R. Frequent wildfires erode tree persistence and alter stand structure and initial composition of a fire‐tolerant sub‐alpine forest. J. Veg. Sci. 28, 1151–1165 (2017).Article 

Google Scholar 
14.Prior, L. D., Williamson, G. J. & Bowman, D. M. J. S. Impact of high-severity fire in a Tasmanian dry eucalypt forest. Aust. J. Bot. 64, 193–205 (2016).Article 

Google Scholar 
15.Bassett, O. D., Prior, L. D., Slijkerman, C. M., Jamieson, D. & Bowman, D. M. J. S. Aerial sowing stopped the loss of alpine ash (Eucalyptus delegatensis) forests burnt by three short-interval fires in the Alpine National Park, Victoria, Australia. For. Ecol. Manag. 342, 39–48 (2015).Article 

Google Scholar 
16.Bowman, D. et al. Wildfires: Australia needs national monitoring agency. Nature 584, 188–191 (2020).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
17.King, A. D., Pitman, A. J., Henley, B. J., Ukkola, A. M. & Brown, J. R. The role of climate variability in Australian drought. Nat. Clim. Change 10, 177–179 (2020).Article 

Google Scholar 
18.Sharples, J. J. et al. Natural hazards in Australia: extreme bushfire. Clim. Change 139, 85–99 (2016).Article 

Google Scholar 
19.Bowman, D. M. J. S., Williamson, G. J., Price, O. F., Ndalila, M. N. & Bradstock, R. A. Australian forests, megafires and the risk of dwindling carbon stocks. Plant, Cell Environ. 44, 347–355 (2020).20.Khaykin, S. et al. The 2019/20 Australian wildfires generated a persistent smoke-charged vortex rising up to 35 km altitude. Commun. Earth Environ. 1, 22 (2020).Article 

Google Scholar 
21.Borchers Arriagada, N. et al. Unprecedented smoke-related health burden associated with the 2019–20 bushfires in eastern Australia. Med. J. Aust. 213, 282–283 (2020).22.Johnston, F. H. et al. Unprecedented health costs of smoke-related PM2.5 from the 2019–20 Australian megafires. Nat. Sustain. 4, 42–47 (2021).Article 

Google Scholar 
23.Ward, M. et al. Impact of 2019–2020 mega-fires on Australian fauna habitat. Nat. Ecol. Evol. 4, 1321–1326 (2020).PubMed 
Article 
PubMed Central 

Google Scholar 
24.Bowman, D. M. J. S., Williamson, G. J., Prior, L. D. & Murphy, B. P. The relative importance of intrinsic and extrinsic factors in the decline of obligate seeder forests. Glob. Ecol. Biogeogr. 25, 1166–1172 (2016).Article 

Google Scholar 
25.Povak, N. A., Kane, V. R., Collins, B. M., Lydersen, J. M. & Kane, J. T. Multi-scaled drivers of severity patterns vary across land ownerships for the 2013 Rim Fire, California. Landsc. Ecol. 35, 293–318 (2020).Article 

Google Scholar 
26.Parks, S. A. et al. High-severity fire: evaluating its key drivers and mapping its probability across western US forests. Environ. Res. Lett. 13, 044037 (2018).Article 

Google Scholar 
27.Fang, L., Yang, J., Zu, J., Li, G. & Zhang, J. Quantifying influences and relative importance of fire weather, topography, and vegetation on fire size and fire severity in a Chinese boreal forest landscape. For. Ecol. Manag. 356, 2–12 (2015).Article 

Google Scholar 
28.Thompson, J. R. & Spies, T. A. Vegetation and weather explain variation in crown damage within a large mixed-severity wildfire. For. Ecol. Manag. 258, 1684–1694 (2009).Article 

Google Scholar 
29.Stephens, S. L. et al. Fire and climate change: conserving seasonally dry forests is still possible. Front. Ecol. Environ. 18, 354–360 (2020).Article 

Google Scholar 
30.Dieleman, C. M. et al. Wildfire combustion and carbon stocks in the southern Canadian boreal forest: implications for a warming world. Glob. Change Biol. 26, 6062–6079 (2020).Article 

Google Scholar 
31.Nolan, R. H. et al. Causes and consequences of eastern Australia’s 2019–20 season of mega‐fires. Glob. Change Biol. 26, 1039–1041 (2020).32.Boer, M. M., Resco de Dios, V. & Bradstock, R. A. Unprecedented burn area of Australian mega forest fires. Nat. Clim. Change 10, 171–172 (2020).Article 

Google Scholar 
33.van Oldenborgh, G. J. et al. Attribution of the Australian bushfire risk to anthropogenic climate change. Nat. Hazards Earth Syst. Sci. 21, 941–960 (2021).Article 

Google Scholar 
34.Adams, M. A., Shadmanroodposhti, M. & Neumann, M. Letter to the Editor. Causes and consequences of Eastern Australia’s 2019‐20 season of mega‐fires: a broader perspective. Glob. Change Biol. 26, 3756–3758 (2020).35.Lindenmayer, D. B. & Taylor, C. New spatial analyses of Australian wildfires highlight the need for new fire, resource, and conservation policies. Proc. Natl Acad. Sci. USA 117, 12481–12485 (2020).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
36.Lindenmayer, D. B., Hobbs, R. J., Likens, G. E., Krebs, C. J. & Banks, S. C. Newly discovered landscape traps produce regime shifts in wet forests. Proc. Natl Acad. Sci. USA 108, 15887–15891 (2011).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
37.Taylor, C., McCarthy, M. A. & Lindenmayer, D. B. Nonlinear effects of stand age on fire severity. Conserv. Lett. 7, 355–370 (2014).Article 

Google Scholar 
38.Collins, L., Griffioen, P., Newell, G. & Mellor, A. The utility of Random Forests for wildfire severity mapping. Remote Sens. Environ. 216, 374–384 (2018).Article 

Google Scholar 
39.Gibson, R., Danaher, T., Hehir, W. & Collins, L. A remote sensing approach to mapping fire severity in south-eastern Australia using sentinel 2 and random forest. Remote Sens. Environ. 240, 111702 (2020).Article 

Google Scholar 
40.Collins, L., Bradstock, R. & Penman, T. Can precipitation influence landscape controls on wildfire severity? A case study within temperate eucalypt forests of south-eastern Australia. Int. J. Wildland Fire 23, 9–20 (2014).Article 

Google Scholar 
41.Price, O. F. & Bradstock, R. A. The efficacy of fuel treatment in mitigating property loss during wildfires: insights from analysis of the severity of the catastrophic fires in 2009 in Victoria, Australia. J. Environ. Manag. 113, 146–157 (2012).Article 

Google Scholar 
42.Storey, M., Price, O. & Tasker, E. The role of weather, past fire and topography in crown fire occurrence in eastern Australia. Int. J. Wildland Fire 25, 1048–1060 (2016).Article 

Google Scholar 
43.Bradstock, R. A., Hammill, K. A., Collins, L. & Price, O. Effects of weather, fuel and terrain on fire severity in topographically diverse landscapes of south-eastern Australia. Landsc. Ecol. 25, 607–619 (2010).Article 

Google Scholar 
44.Taylor, C., Blanchard, W. & Lindenmayer, D. B. Does forest thinning reduce fire severity in Australian eucalypt forests? Conserv. Lett. https://doi.org/10.1111/conl.12766 (2020).45.Lydersen, J. M. et al. Evidence of fuels management and fire weather influencing fire severity in an extreme fire event. Ecol. Appl. 27, 2013–2030 (2017).PubMed 
Article 
PubMed Central 

Google Scholar 
46.Gómez-González, S., Ojeda, F. & Fernandes, P. M. Portugal and Chile: longing for sustainable forestry while rising from the ashes. Environ. Sci. Policy 81, 104–107 (2018).Article 

Google Scholar 
47.Bowman, D. M. J. S. et al. Human–environmental drivers and impacts of the globally extreme 2017 Chilean fires. Ambio 48, 350–362 (2019).PubMed 
Article 
PubMed Central 

Google Scholar 
48.Jackson, W. Fire, air, water and earth–an elemental ecology of Tasmania. Proc. Ecol. Soc. Aust. 3, 9–16 (1968).
Google Scholar 
49.Tolhurst, K. G. & McCarthy, G. Effect of prescribed burning on wildfire severity: a landscape-scale case study from the 2003 fires in Victoria. Aust. For. 79, 1–14 (2016).Article 

Google Scholar 
50.Gammage, B. The Biggest Estate on Earth: How Aborigines Made Australia (Allen & Unwin, 2011).51.Dargavel, J. Views and perspectives: why does Australia have ‘forest wars’? Int. Rev. Environ. Hist. 4, 33–51 (2018).Article 

Google Scholar 
52.Kanowski, P. J. Australia’s forests: contested past, tenure-driven present, uncertain future. For. Policy Econ. 77, 56–68 (2017).Article 

Google Scholar 
53.Australian Forest and Wood Products Statistics Mar-Jun 2019 (Australian Bureau of Agricultural and Resource Economics and Sciences, 2019).54.Ferguson, I. Australian plantations: mixed signals ahead. Int. For. Rev. 16, 160–171 (2014).
Google Scholar 
55.Raison, R. & Squire, R. Forest Management in Australia: Implications for Carbon Budgets Technical Report 32 (Australian Greenhouse Office, 2008).56.Proctor, E. & McCarthy, G. Changes in fuel hazard following thinning operations in mixed-species forests in East Gippsland, Victoria. Aust. For. 78, 195–206 (2015).Article 

Google Scholar 
57.NSW Regional Forest Agreements Assessment of Matters Pertaining to Renewal of Regional Forest Agreements (NSW Department of Primary Industries, 2018).58.Evans, J. spatialEco_. R package version 1.3-1 https://github.com/jeffreyevans/spatialEco (2020).59.Farr, T. G. et al. The shuttle radar topography mission. Rev. Geophys. https://doi.org/10.1029/2005RG000183 (2007).60.Dowdy, A. J. Climatological variability of fire weather in Australia. J. Appl. Meteorol. Climatol. 57, 221–234 (2018).Article 

Google Scholar 
61.Hodges, J. S. & Reich, B. J. Adding spatially-correlated errors can mess up the fixed effect you love. Am. Stat. 64, 325–334 (2010).Article 

Google Scholar 
62.Rabus, B., Eineder, M., Roth, A. & Bamler, R. The shuttle radar topography mission—a new class of digital elevation models acquired by spaceborne radar. ISPRS J. Photogramm. Remote Sens. 57, 241–262 (2003).Article 

Google Scholar 
63.Kuhn, M. et al. caret: Classification and regression training. R package version 6.0-77 (2018).64.De Reu, J. et al. Application of the topographic position index to heterogeneous landscapes. Geomorphology 186, 39–49 (2013).Article 

Google Scholar  More

1.Cerdà, A., González Pelayo, Ó., Pereira, P., Novara, A., Iserloh, T. et al. The wildgeographer avatar shows how to measure soil erosion rates by means of a rainfall simulator. In Geophysical Research (2015).2.Fu, B. J. Soil erosion and its control in the Loess Plateau of China. Soil Use Manag. 5, 76–81 (1989).Article 

Google Scholar 
3.Lal, R. Soil carbon sequestration impacts on global climate change and food security. Science 304(5677), 1623–1627 (2004).ADS 
CAS 
Article 

Google Scholar 
4.Govers, G., Van Oost, K. & Wang, Z. Scratching the critical zone: the global footprint of agricultural soil erosion. Procedia Earth Planet. Sci. 10, 313–318 (2014).ADS 
Article 

Google Scholar 
5.Zhou, P., Wen, A., Zhang, X. & He, X. Soil conservation and sustainable eco-environment in the Loess Plateau of China. Environ. Earth Sci. 68(3), 633–639 (2012).
Google Scholar 
6.Aldaood, A., Bouasker, M. & Al-Mukhtar, M. Soil–water characteristic curve of lime treated gypseous soil. Appl. Clay Sci. 102, 128–138 (2014).CAS 
Article 

Google Scholar 
7.Satyanaga, A., Rahardjo, H., Leong, E. C. & Wang, J. Y. Water characteristic curve of soil with bimodal grain-size distribution. Comput. Geotech. 48(4), 51–61 (2013).Article 

Google Scholar 
8.Li, X., Li, J. H. & Zhang, L. M. Predicting bimodal soil–water characteristic curves and permeability functions using physically based parameters. Comput. Geotech. 57(4), 85–96 (2014).MathSciNet 
CAS 
Article 

Google Scholar 
9.Breda, N., Huc, R., Granier, A. & Dreyer, E. Temperate forest trees and stands under severe drought: a review of ecophysiological responses, adaptation processes and long-term consequences. Ann. For. Sci. 63, 625–644 (2006).Article 

Google Scholar 
10.Zhao, J., Lin, L., Yang, K., Liu, Q. & Qian, G. Influences of land use on water quality in a reticular river network area: a case study in Shanghai, China. Landsc. Urban Plan. 137, 20–29 (2015).Article 

Google Scholar 
11.McIntosh, J. C. & Horne, D. J. Causes of repellency: I. The nature of the hydrophobic compounds found in a New Zealand development sequence of yellow-brown sands. In Proceedings of the 2nd National Water Repellency Workshop, 1–5 August, Perth, Western Australia, 8–12 (1994).12.de Jonge, L. W., Moldrup, P. & Jacobsen, O. H. Soil-water content dependency of water repellency in soils: effect of crop type, soil management, and physical–chemical parameters. Soil Sci. 172, 577–588 (2007).ADS 
Article 

Google Scholar 
13.Hallett, P. D., Ritz, K. & Wheatley, R. E. Microbial derived water repellency in golf course soil. Int. Turfgrass Soc. Res. J. 9, 518–524 (2001).
Google Scholar 
14.Caravaca, F., Masciandaro, G. & Ceccanti, B. Land use in relation to soil chemical and biochemical properties in a semiarid Mediterranean environment. Soil Till. Res. 68(1), 23–30 (2002).Article 

Google Scholar 
15.Li, H., Yan, F. C., Jiao, J. Y., Tang, B. Z. & Zhang, Y. F. Soil water availability and holding capacity of different vegetation types in hilly-gullied region of the loess plateau. Acta Ecol. Sin. 38(11) (2018).16.Ritchie, J. T. Soil water availability. Plant Soil 58(58), 327–338 (1981).Article 

Google Scholar 
17.Wan, S., Norby, R. J., Ledford, J. & Weltzin, J. F. Responses of soil respiration to elevated CO2, air warming, and changing soil water availability in a model old-field grassland. Glob. Change Biol. 13(11), 2411–2424 (2007).ADS 
Article 

Google Scholar 
18.An, S. et al. Soil quality degradation processes along a deforestation chronosequence in the Ziwuling area, China. CATENA 75(3), 248–256 (2008).Article 

Google Scholar 
19.Wang, K. B., Shao, R. X. & Shangguan, Z. P. Changes in species richness and community productivity during succession on the Loess Plateau (China). Pol. J. Ecol. 58(3), 501–510 (2010).
Google Scholar 
20.Wang, Y., Shao, M. & Shao, H. A preliminary investigation of the dynamic characteristics of dried soil layers on the Loess Plateau of China. J. Hydrol. 381(1–2), 9–17 (2010).ADS 
Article 

Google Scholar 
21.Reatto, A., Silva, E. M. D., Bruand, A., Martins, E. S. & Lima, J. E. F. W. Validity of the centrifuge method for determining the water retention properties of tropical soils. Soil Sci. Soc. Am. J. 72(6), 1547–1553 (2008).ADS 
CAS 
Article 

Google Scholar 
22.Jia, G. M., Cao, J., Wang, C. Y. & Wang, G. Microbial biomass and nutrients in soil at the different stages of secondary forest succession in Ziwuling northwest China. For. Ecol. Manag. 217, 117–125 (2005).Article 

Google Scholar 
23.Ghanbarian-Alavijeh, B. & Millán, H. The relationship between surface fractal dimension and soil water content at permanent wilting point. Geoderma 151(3–4), 224–232 (2009).ADS 
Article 

Google Scholar 
24.Wang, M. B., Chai, B. F., Li, H. J. & Feng, C. P. Soil water holding capacity and soil available water in plantations in the loess region. Sci. Silvae Sin. 35(2), 7–14 (1999).
Google Scholar 
25.Yang, L., Wei, W., Chen, L. D. & Mo, B. Response of deep soil moisture to land use and afforestation in the semi-arid Loess Plateau, China. J. Hydrol. 475(6), 111–122 (2012).ADS 
Article 

Google Scholar 
26.Huang, J. H., Liao, Y. C., Gao, M. S. & Yin, R. J. Effects of tillage and mulching on orchard soil moisture content and temperature in Loess Plateau. Chin. J. Appl. Ecol. 20(11), 2652–2658 (2009).
Google Scholar 
27.Borken, W., Savage, K., Davidson, E. A. & Trumbore, S. E. Effects experimental drought on soil respiration and radiocarbon efflux from a temperate forest soil. Glob. Change Biol. 12, 177–193 (2006).ADS 
Article 

Google Scholar 
28.Scott-Denton, L. E., Rosenstiel, T. N. & Monson, R. K. Differential controls by climate and substrate over the heterotrophic and rhizospheric components of soil respiration. Glob. Change Biol. 12(12), 205–216 (2006).ADS 
Article 

Google Scholar 
29.Zhang, Y. W., Deng, L., Yan, W. M. & Shangguan, Z. P. Interaction of soil water storage dynamics and long-term natural vegetation succession on the Loess Plateau, China. CATENA 137, 52–60 (2016).Article 

Google Scholar 
30.Zhao, S. W., Zhao, Y. G. & Wu, J. S. Quantitative analysis of soil pores under natural vegetation successions on the Loess Plateau. Sci. China Earth Sci. 53, 617–625 (2010).ADS 
CAS 
Article 

Google Scholar 
31.Udawatta, R. P. & Anderson, S. H. Ct-measured pore characteristics of surface and subsurface soils influenced by agroforestry and grass buffers. Geoderma 145(3–4), 381–389 (2008).ADS 
Article 

Google Scholar 
32.Honda, E. A. & Durigan, G. Woody encroachment and its consequences on hydrological processes in the Savannah. Philos. Trans. R. Soc. B 371(1703), 20150313 (2016).Article 

Google Scholar 
33.Zhang, Y. W. & Shangguan, Z. P. Interaction of soil water storage and stoichiometrical characteristics in the long-term natural vegetation restoration on the Loess Plateau. Ecol. Eng. 116, 7–13 (2018).Article 

Google Scholar 
34.Wang, Z. H. et al. Effects of plant species diversity on soil conservation and stability in the secondary succession phases of a semi-humid evergreen broadleaf forest in China. J. Soil Water Conserv. 67, 311–320 (2012).Article 

Google Scholar 
35.Yang, L., Wei, W., Chen, L. D. & Wang, J. L. Response of temporal variation of soil moisture to vegetation restoration in semi-arid Loess Plateau, China. CATENA 115, 123–133 (2014).Article 

Google Scholar 
36.Wang, L., Mu, Y., Zhang, Q. F. & Jia, Z. K. Effects of vegetation restoration on soil physical properties in the wind–water erosion region of the northern Loess Plateau of China. Clean: Soil, Air, Water 40(1), 7–15 (2012).
Google Scholar 
37.Deng, L., Wang, K. B., Chen, M. L., Shangguan, Z. P. & Sweeney, S. Soil organic carbon storage capacity positively related to forest succession on the Loess Plateau, China. CATENA 110, 1–7 (2013).CAS 
Article 

Google Scholar 
38.Zhang, Y. W. & Shangguan, Z. P. The coupling interaction of soil water and organic carbon storage in the long vegetation restoration on the Loess Plateau. Ecol. Eng. 91, 574–581 (2016).Article 

Google Scholar 
39.Luo, Y. Q. et al. Progressive nitrogen limitation of ecosystem responses to rising atmospheric carbon dioxide. Bioscience 54, 731–739 (2004).Article 

Google Scholar  More

We recorded the songs of 160 zebra finch tutor–pupil pairs (68 tutors and 160 pupils; 228 birds overall) at the Rockefeller University Field Research Center colony, which consisted of over 800 birds during the 1-year period of recording. Of the 160 pupils, 130 pupils were housed with their biological parents, and 30 pupils with foster parents. We also analyzed song imitation across three generations including 14 grand-tutors and 35 grand-pupils. All birds were housed in individual breeding cages with parents (either biological or foster) and other offspring, and kept visually isolated from adjacent breeding cages. With this social regimen, we found no evidence of song imitation across families (Supplementary Fig. 1). From each bird, we recorded undirected songs (produced in isolation) for over a week to obtain a sample of at least 1000 song syllables per bird. Directed songs to females were also recorded, but not analyzed for this study.Imitation outcome varied across familiesWe first measured similarity27 between tutor and pupil songs based on acoustic features (pitch, frequency modulation, Wiener entropy, and spectral continuity)28. We observed considerable variability in the distribution of song similarities between pupils and their individual tutors (mean = 69%; range 20–100%; CV = 0.28, Fig. 1a). To test for family influence, we identified 24 families that had multiple clutches with males, calculated the mean song similarity between pupils and their tutors of each clutch, which allowed us to normalize out the effect of song convergence between siblings20. We then calculated the coefficients of variance across clutches within families and compared it to the coefficient of variance across families (Fig. 1b). We found that imitation similarity was much more variable across families than within families (Kruskal–Wallis chi-squared = 44.727, df = 23, p-value = 0.006).Fig. 1: Distribution of song similarity between pupils and their tutors.a Histogram of song similarities between 160 pupils and their tutors. b Analysis of variance in song similarity between and within families. Data include 24 families with more than one clutch with males. Similarity scores were averaged within clutch members and coefficient of variance (CV = 0.14 ± 0.02) of similarity scores were calculated across clutches. CV of the same data (averaged within clutches) across families = 0.24 is presented as a dotted line. Source data for this figure is in Supplementary Data File 1.Full size imageIn certain families, across clutches, song imitation tended to be almost exclusively accurate (top quartile), in some modest (middle quartile), and others generally poor (Fig. 2a). To assess whether this variance in imitation outcome was genetic, we compared song imitation between biological and foster pupils. Foster pupils imitated their tutor as well as biological ones (biological similarity: 68.2 ± 1.7%, n = 130; foster similarity: 70.0 ± 3.6%, n = 30, mean ± S.E.M. hereafter). Therefore, the variability we observed in imitation outcomes across families cannot be explained by genetic variability. Instead, we noted that variability in imitation among pupils appeared to be associated with tutor song structure. For example, tutor Aq12 had a very simple song with one syllable-type containing two notes and none of his pupils imitated this syllable or song accurately. Instead, some pupils introduced apparently novel syllable types not found in the tutor in developing their own songs (Fig. 2b). In contrast, tutor DG1 had a more complex song, with five syllable types containing six notes, and all of his pupils imitated the syllables and the sequence much more accurately, with little to no introduction of novel syllables (Fig. 2c). In both cases, pupils still produced their syllables in repeated song motifs of 2–6 syllable types, as is typical of zebra finches (Fig. 2b, c). This suggested to us that pupils might more accurately imitate tutor songs that are rich in acoustic structure (i.e., acoustically diverse), while improvising upon impoverished tutor songs.Fig. 2: Imitation outcome varies across families.a 24 song tutoring lineages. All tutors had pupils in more than a single clutch. Each node represents one individual animal. Node shape represents pupils from the same clutch. Tutor nodes are presented on the bottom and pupil nodes on the top. Similarity scores are presented as quartiles (green for best imitations and red for poorest). Lineages are sorted according to the mean similarity between tutor and pupils from highest (top) to lowest (bottom). b, c Examples of song imitations from tutor AQ12 with a low similarity family (b) and from tutor DG1 with a high similarity family (c). Imitation outcomes are presented as percent acoustic similarity estimates on each sonogram. Red bars outline the repeated song motifs of the tutors. Source data for this figure is in Supplementary Data File 1.Full size imageSyllable-type diversity is not correlated between tutor and pupil songsIf this impoverished tutor song hypothesis were true, we would expect to find that as tutor syllable diversity decreases, pupil’ imitation similarity also decreases; conversely, we would expect to see biases in the correlation between tutor syllable diversity and pupil syllable diversity at extreme ranges of tutor diversity. To test this hypothesis quantitatively, we sought a measure of syllable acoustic and syntax diversity. We selected a random group of 80 adult tutor–pupil pairs, and segmented their songs into syllable units using an amplitude threshold27. Song syllables were automatically clustered into types based on their acoustic features (Fig. 3a, b)27. We then calculated the relative frequency (abundance) of each syllable-type and used Shannon information entropy27 to measure syllable acoustic diversity produced by each bird. Specifically, for each bird’s song, we calculated the proportion ({p}_{i}) of syllables produced for each syllable-type i, and computed entropy as (-sum {p}_{i}({{{log}}}_{2}({p}_{i}))). The measure weighs each vocal element (syllable) by its abundance, and presents the entropy (diversity) of the distribution in units of bits. We used the same Shannon information measure to also evaluate syllable transition diversity (song-syntax entropy29). The Shannon information entropy has limited bearing on capturing combinatorial complexity30, but it is a better estimate of diversity compared to just counting syllable types because it considers the frequencies (abundances) of each type. The more syllable types produced, and the more even their abundances are, the higher the entropy.Fig. 3: Syllable-type diversity.a Example sonograms of a tutor–pupil song pair. Syllable types are color-coded by lines above them. Color lines above each syllable indicate clusters computed separately for tutor and pupil in b. Note that syllable types are bird specific and color codes have no correspondence between tutor and pupil, e.g., green, yellow, and black labeled syllable types in tutor song merged into a single type (yellow labeled) in pupil’s song. b 2D scatter plots of syllable acoustic features: duration versus mean pitch, mean frequency modulation (FM), or mean Wiener entropy (a measure of the width of the power spectrum). The color of each marker indicates its computed syllable-type (type = cluster in feature space). Colors of clusters correspond to syllable-type colors shown in a. c Histogram of syllable-type diversity, pooled across all birds. d Regression analysis between tutor and pupil syllable-type diversity, showing no significant correlation for pupils with high or low imitation similarity of their tutors. e, f Tutor syllable diversity is not correlated with pupil song imitation similarity (e), or influence of tutors on pupils (f). g–k Examples of five tutor–pupil pairs with syllable recombination, namely merging in pupil songs. Source data for this figure is in Supplementary Data File 1.Full size imageThe distribution of syllable-type diversity of songs in the population was asymmetric, with most songs in the range of 2.5–3 bits and a left tail of rare songs with low syllable diversity (Fig. 3c). Surprisingly, there was no statistically significant correlation between tutor and pupil syllable diversity (R2 = 0.079, NS). Looking separately at pupils who imitated above (and below) average showed no correlations either (Fig. 3d). Further, there was no correlation between tutor syllable diversity and acoustic similarity between tutor and pupil songs (Fig. 3e). To better estimate how tutor syllable diversity may affect the cultural transmission, we calculated song acoustic similarity in reverse, from pupil to tutor. We call this a measure of “influence” because it tells us how much of the pupil’s song is influenced by the tutor. However, influence in pupils was not significantly correlated with tutor song (Fig. 3f). Near zero correlations were also observed for song-syntax (bigram) transitions between pairs of syllable types (Supplementary Fig. 1). In sum, our syllable-type diversity measure failed to capture any aspect of song learning, nullifying all our attempts to evaluate our impoverished tutor song hypothesis.Half of the pupils recombine syllablesPuzzled by the lack of even a weak correlation between tutor and pupil syllable and syntax diversity, we examined cases of most accurate imitation. We found frequent inconsistencies, as is typical of zebra finches in the boundaries of corresponding syllables in the songs of tutors and their pupils, even in cases of accurate imitation. This was not primarily due to measurement (segmentation) errors, but because pupils often modified or recombined the units they imitated (Fig. 3g). We assessed a lower bound estimate of similarity in the syllable boundaries of tutor and pupil songs, restricting the analysis to those syllables whose acoustic structure was clearly and fully imitated (either as a single unit or in parts) by the pupil (examples in Fig. 3g–k). With this strict criterion, analysis of syllable imitations in 33 randomly selected tutor–pupil pairs revealed modification of syllable boundaries in 47 cases (22%) of the copied syllables. Overall, 54% (18/33) of the pupils showed at least one case of altering syllables units. Interestingly, all 47 cases were of merging tutor syllables, rather than splitting. However, splitting might be more difficult to detect, and if so, our analyses would be an underestimate of the magnitude of syllable recombination (see “Methods” section).Vocal state measures capture balanced imitationGiven the extent of syllable recombination, we next sought an alternative quantitative measure that captures acoustic diversity at the sub-syllabic level, which would be, by design, insensitive to syllable recombination. For each of the 160 tutor–pupil pairs, we calculated continuously (in 10 ms FFT windows excluding silences, but without segmentation) three acoustic feature vectors: pitch, Wiener entropy (width of power spectrum), and frequency modulation28. Histograms of these features for all birdsongs in our sample reveal several concentrations, and we used the contours of these concentrations to partition the entire acoustic space of the songs into 10 regions. To visualize these concentrations, we present 2D slices of the feature space according to four peaks in the distribution of pitch (Fig. 4a), which we labeled very low, low, medium, and high. These four slices show distinct concentrations of the 10 regions, that we will call vocal states (Fig. 4b). The two concentrations in the highest and lowest pitch regions consisted of down-modulated and up-modulated sounds, respectively (vocal states 1 and 2, for lowest pitch, and 9 and 10 for highest pitch). The two central pitch regions (low and medium) consisted of similar types of vocal states, and two additional states (4 and 7) centered at zero frequency modulation represent non-modulated harmonic sounds. With the vocal states of the population categorized, we can consider each song as a long sequence of vocal states, calculated in small (10 ms) time windows. We next analyzed the distribution of vocal states, by calculating the relative abundances of sounds within each vocal state for each bird.Fig. 4: Vocal states and diversity in zebra finch songs.a Histogram of pitch, calculated in 10 ms windows and pooled for all songs31. Shadings show partitioning into four regions according to contours of the pitch distribution. b Two-dimensional heatmaps of frequency modulation and Wiener entropy for each of the four-pitch regions. Red circles outline 10 clusters around which vocal states are defined. c Histogram of song diversity for all male birds recorded. d Song diversity in tutor songs versus pupil songs. Colors show R2 separately for high and low similarity birds. e Tutor song diversities vs. similarity with pupil songs (R2 = 0.08, t = 1.9, Linear mixed-effects model NS). f Tutor song diversities vs. the influence of tutor song on pupils (R2 = 0.25, Linear mixed-effects model t = 4.8, p = 4.2e−6). Vertically aligned markers are often birds from the same lineage. The trend remains significant after removing the lowest diversity families (ABCDEF will give us 100% imitation similarity because all of the tutor’s sounds are present in pupil’s song, but only 50% influence because half of the pupil’s song is improvised. Indeed, for tutors with high song diversity, the diversity of pupil’ songs is centered on the diagonal (identity) line (Fig. 4d). However, for tutors with low song diversity, pupil song diversity, in most cases, above the diagonal (Fig. 4d). For example, out of 34 tutor songs with diversity below 3, only 5 pupil songs are below the diagonal. That is, pupils of tutors with low song diversity often imitated them, but were less influenced by them: they often made additions that increase song diversity. These “low influence” songs did not resemble neighboring birds’ songs (Supplementary Note 1). We, therefore, suspect that these additions are improvisations, namely they are likely to be either modified versions of tutor song elements, or innate syllable types.Assuming a natural trend to develop high-diversity songs either via imitation or improvisation, we wondered why songs of low diversity were not rarer in our colony. We tested which factors may sustain songs of low diversity across generations and found that pupils that imitated poorly, regardless of tutor song diversity, tended to have low-diversity songs (Fig. 4g, R2 = 0.20, t = 5.7, p = 6.2e−8). To directly test for interaction between imitation accuracy and song diversity, we ran a linear mixed-effect model to explain pupil song diversity with two fixed effects: the diversity of the tutor song, and the acoustic similarity to the tutor song (how much of it was copied). Results confirmed that both factors contribute about equally to pupil song diversity (imitation similarity: t = 5.0, p = 1.4e−6; tutor song diversity: t = 4.6, p = 7.9e−6).In sum, although our syllable diversity measure failed to capture any relationship with song imitation, bypassing syllable recombination by measuring song diversity based on vocal states (without segmentation) revealed two effects: First, low diversity in a tutor’s song was not associated with lower imitation similarity in the pupil but with lower influence on the pupil, indicating a tendency in pupils to increase song diversity, which we call “balanced imitation”. Second, low diversity in a pupil’s song (but not in tutor song) is associated with poor imitation similarity in the pupil. Together these effects can explain the stable polymorphism in song diversity across generations: on the one hand, pupils tend to increase song diversity when tutored by a low-diversity song model, but on the other hand, poor imitation is associated with a decrease of song diversity in pupils’ songs. Consistent with this interpretation, when we plotted song diversity of each tutor against the mean song diversity of all of his pupils, the mean song diversity in pupils of low-diversity (below median) tutors was often higher than that of their tutors, and vice versa (Fig. 4h). That is, despite the overall positive correlation between tutor and pupil song diversity, we see frequent reversals such that a large proportion (42%) of pupils with low song diversity had tutors with high (above median) song diversity, and vice versa.Balanced imitation across multiple generationsWe further explored reversals across multiple generations, and analyzed 14 family branches, where we had song imitation data across two generations of pupils. We found that in the families where the first-generation pupils imitated poorly, there was often some recovery in imitation accuracy in the second-generation, the grand-pupils (Fig. 5a). For example, in the two lineages (HP10 and DG4) with the greatest number of first-generation pupils that imitated poorly, all of the second-generation (grand) pupils imitated the song of their tutor more accurately than the tutor’s imitation of the grand tutor. Sonograms revealed that, in both lineages, the grand-tutor songs were unbalanced: Tutor HP10 had a very high-pitched song (Fig. 5b), whereas tutor DG4’s song included numerous harmonic stacks (Fig. 5c). In both cases, their pupils developed songs that appear to be more acoustically “balanced,” and ones that the grand-pupils imitated accurately (Fig. 5b, c). In other cases, however, low similarity was simply due to partial imitation, e.g., in the lineage (LB12), where the song imitation became worse because a grand–pupil dropped a syllable during imitation (Fig. 5d). These findings suggest that grand-pupils of impoverished-song grand tutors imitate some elements from the deficient songs of their tutors, but they also further “balance” them, thus increasing the diversity of their songs.Fig. 5: Song diversity across generations.a Song similarity across two generations of pupils (colors represent quartiles (as in Fig. 1a)) in 14 family lineages. b An example from lineage HP10 showing a transition from poor imitation in a first-generation pupil to accurate imitation in a grand pupil. c Same as b for lineage DG4. d A counter example in lineage LB12, where the grand pupil imitated poorly. Source data for this figure is in Supplementary Data File 1.Full size imageBalanced imitation of vocal state abundancesOur measures up to now summarize the distribution of vocal states within a song. We next looked at each vocal state separately and measured how frequencies (abundances) of vocal states are imitated. In prior studies, we noted that vocal imitation in zebra finches is inversely related to model abundance. That is, too much exposure to a tutored song could inhibit learning31. Here we test if this is the case also for abundances of vocal states within a song.We partitioned the vocal state data into quartiles based on the overall acoustic similarity between tutor and pupil songs. For each tutor–pupil pair, in each quartile, we then plotted the relative abundances of all 10 corresponding vocal states in the tutor’s song versus his pupil’s song (Fig. 6a–d). We found that relative abundances of all 10 states were correlated, for each quartile. As expected, tutor–pupil vocal state abundances were more strongly associated when imitations were accurate; for example, the residual coefficient of determination was much higher in the top similarity quartile, explaining about 35% of the variance in cases of highest song similarity (Fig. 6a), and only about 9% of the variance in the bottom quartile (Fig. 6d). We noted that in all quartiles, the slope of the correlation was less than one (Fig. 6a–d), meaning that when tutor’s vocal state was low in abundance, pupil’s vocal states tended to be higher (above the diagonal) and vice versa.Fig. 6: Imitation of vocal state abundances.a–d Scatter plots of tutor vs. corresponding pupil vocal state abundances according to quartiles of song similarities. Note that each bird is represented by 10 markers, which are not statistically independent. The residual correlations were computed after removing trends with bird identities included as random factors. Dashed lines are identity, slope = 1. Colored lines are regression of the data. e Same data as in a–d combined, comparing vocal states abundances  > 20% in tutor vs. pupil songs. f Median imitation gains for all state abundances, according to imitation quartiles. Gain of 1 indicates no bias, gain of 2 indicates doubling of abundance, and gain of 0.5 halving. Y axis is log-scale. g–i Examples of song diversity balancing. We simplified the 10 vocal states into 4 groups: yellow for high pitch states 9–10; mustard for medium pitch, high entropy states 6 and 8; light blue for non-modulated states 4 and 7; and dark blue for the rest 1, 2, 3, and 5. In i, we present two generations of pupils. Note the more uniform pie charts in pupils compared to their tutors. j, Vocal state abundances in biological tutors vs. pupils’ songs. k Vocal state abundances in foster tutors vs. pupils’ songs. l Vocal state abundances in fostered pupils vs. their biological fathers, who did not raise them, which did not raise them. Dashed lines are identity, slope = 1. Red lines are regression of the data. Source data for this figure is in Supplementary Data File 1.Full size imageWe next tested for statistical significance of this bias across the entire data set. Our null hypothesis is that when the abundance of a vocal state in the tutor’s song is high, his pupil is not more likely than chance to deviate from the model in a manner that “balances” his song. In other words, if deviations (imitation “errors”) are random, then the likelihood of deviations (errors) to increase or decrease song diversity should be determined by the overall distribution of errors in our sample. In a previous study2, some of us presented evidence that imitation of isolated tutors is biased: syllables with high abundance in abnormal isolate tutor song ( >20%) were often less abundant is pupil’s songs. Using the same 20% threshold we found that the distribution of tutor vs. pupil vocal state abundances is asymmetric (Fig. 6e): when tutor’s vocal state abundance is above 20%, about 14% of corresponding pupil’s states are above the diagonal (hence 86% of the errors increase song diversity). But looking in reverse, we found that when a pupil’s vocal state is above 20%, a higher proportion of corresponding tutor’s states (23%) are to the right of the diagonal. To overcome dependencies between vocal states, we treated each tutor–pupil pair as a statistic. We randomly shuffled the direction tutor- >pupil vs. pupil- >tutor (without breaking the pairs) to obtain a random distribution of biases. We found that the observed bias to increase song diversity (namely in the direction that decreases the abundance of vocal states that are already of high abundance) is higher than expected by chance (bootstrap direct p-value = 0.032).We wondered if this bias is stronger in cases of poor imitation, due to the inclusion of non-tutor syllables (via improvisation or innate vocalization). To evaluate if this was the case, we divided the tutors’ vocal states into 0.1 abundance bins, then calculated the median abundance of pupil vocal states for each bin. For each bin, we calculated the abundance ratio for that median. For example, if at the window centered at 0.1 tutor abundance, the median pupil vocal state abundance was 0.2, then the gain ratio would be 2. A gain value of 1 (y axis in Fig. 6f) represents the identical abundance of all 10 vocal states in pupil and tutor. A gain value of 2 indicates a doubling of abundances in the pupil (amplification), and a value of 0.5 halving (attenuation). Interestingly, the gain-loss curves have similar shapes and magnitude across all four quartile groups (Fig. 6f). In all cases, a gain of 1 (where abundance tends to be identical across pupils and their tutors), was at 11–12% abundance, which is fairly close to the center of the distribution (=10%, since we have 10 vocal states). These findings suggest that the regression we noted is not an entirely random effect. For example, in Q1, where the mean similarity is 93%, we see that when tutor state abundance is above 0.2, the corresponding pupil abundance is lower in 10 out of 11 cases (Fig. 6a). In all these cases, the corresponding vocal sounds were imitated, but produced either less often, or with biased features, by the pupil.To visually compare vocal state abundances in tutor vs pupil songs, we reduced the ten vocal states into four color codes, and graphed them along with the sonograms of each bird (Fig. 6g–i). In cases where the tutors’ songs included many high-pitched vocalizations (vocal states 9 and 10), their pupils imitated, but lowered the pitch, thereby decreasing the abundance of those states (Fig. 6g, h). In another example, where the tutor’s song had a high abundance of harmonic stacks (states 4 and 7), their pupil imitated only a subset of these sounds (Fig. 6i). In turn, in the following generation, the pupil’s pupil further differentiated his song to include more balanced vocal states (Fig. 6i). Taken together, song imitation appears to be highly sensitive to the relative abundances of vocal states, suggesting a balancing mechanism that prevents song diversity from becoming too low, perhaps independently of imitation.Finally, we asked whether fostered pupils imitate their tutor’s song vocal states as accurately as biological pupils. Analysis at the level of vocal states allowed us to compare how abundances of vocal states are influenced by foster vs. biological fathers. For reference, imitation of vocal state abundances between the 130 biological pupils and their fathers had a residual R2 = 0.16 (Fig. 6j; t = 5.9, p = 3.9e−09). The 30 foster pupils relative to their foster fathers had a similar R2 = 0.19 (Fig. 6k; t = 2.5, p = 0.01). This is supported by a near-zero correlation between fostered pupils and their biological fathers (Fig. 6l, residual R2 = 0.01, t = 0.46, NS). Therefore, the similarities we observed in vocal state abundances between tutors and their pupils reflect learning with no detectable genetic effect at this level of analysis.How balanced imitation constrains distributions of song featuresWe first tested if abundances of specific vocal states are similar across low-diversity and high-diversity tutor songs. We pooled together songs from tutors that had the lowest diversity (bottom quartile) and calculated the diversity of their “pooled song”. We found that the diversity increased from a mean of 2.99 bits to 3.17 bits, which is similar to the mean diversity in the top quartile (mean = 3.16 bits) but lower than the pooled diversity of the top quartile (=3.27 bits). This outcome indicates that the distribution of vocal states pooled across low-diversity songs is fairly broad, but not as broad as that across songs of high diversity. The distribution of abundances of pooled vocal states (Fig. 7a) explained this difference: As opposed to the nearly flat distribution of vocal state abundances in the high-diversity songs, low-diversity songs tend to have a higher proportion of states 9 and 10, which correspond to high pitch sounds. This is interesting because, in this respect, the low-diversity songs are structurally similar to isolate songs, which are often of higher pitch32. As expected, comparing top and bottom quartiles of influence on the pupil show a similar outcome (Fig. 7b). This outcome suggests that mean song features of low and high influence songs should differ. Further, the variance should also differ: High-diversity songs by definition cannot be extreme in their mean feature values. Low-diversity songs can, in principle, have average features that are close to the population mean, but are more likely to have extreme mean feature values. For example, a song containing mostly high-pitched sounds is both low diversity and extreme in its mean pitch (see for example tutor HP10 in Fig. 5b).Fig. 7: Song diversity versus imitation.a Vocal state abundances in pupils pooled over birds with lowest (bottom quartile) song diversity (dotted line) vs. top quartile (solid line). b Same as a for the bottom (red) and top (green) quartiles of tutor song influence. c–e Mean tutor’s song features versus pupil’s song features for pitch (c), frequency modulation (d), and Wiener entropy (e) for the top influences (green dots, top quartile) and for bottom influences (red dots, bottom quartile). Plotted at the bottom are histogram lines of tutor features for top and bottom quartiles. f–h Box plot distribution of mean song features in four colonies for pitch (f), frequency modulation (g), and Wiener entropy (h). Each marker represents the mean value for one bird. Green shaded areas correspond to top influence feature ranges in colony RU 2019 (this study), whereas red shaded areas correspond to bottom influence feature ranges in colony RU 2019 (n = 149 birds). In the box plots themselves, the red line is the median; Orange fill are the upper and lower quartiles; Blue fill is the minima and maxima. About 20% of the RU 2019 colony are descendants from the RU 2002 colony (Rockefeller Nottebohm Lab; n = 42 birds). The remainder of the 2019 colony originated from Duke University. Colony 3 is from the University of Southern California (Bottjer Lab; n = 48) and Colony 4 is from Cornell University (Regan Lab; n = 77). Source data for this figure is in Supplementary Data File 1.Full size imageWe asked whether we can predict imitation outcomes based on the mean features of a tutor song. If songs of low diversity were culturally transmitted less than high-diversity songs, then songs with extreme mean features—which are typically of low diversity—should be transmitted less. To evaluate this, we plotted the mean pitch of tutor songs against the pitch of their pupil’s songs. Indeed, the distribution of mean song pitch was tighter for the top quartile of tutor–pupil song imitation (Fig. 7c). For example, all tutor songs with a mean pitch above 2000 Hz were of low influence (Fig. 7c, histogram red symbols); these extreme songs were also of low diversity. A similar effect can be seen in Wiener entropy (Fig. 7d) and frequency modulation (Fig. 7e): in both cases the distributions were broader for low-diversity songs. Further, for mean pitch, top influence (green line) is equal or higher than low influence (red line) between 795 Hz and 1885Hz (Fig. 7c). Bottom influence is higher between 1885 and 3000 Hz (red line above green line, Fig. 7c).We superimposed these empirically determined pitch intervals (for top and bottom influence) on ranges of mean song pitches obtained in a database of four zebra finch colonies including the current one, and shaded the intervals values green (presumably top influence) and red (presumably low influence; Fig. 7f). We then did the same for frequency modulation (Fig. 7g), and Weiner entropy (Fig. 7h). Across the colonies, the distribution of mean song features was to a large extent confined within the range of high influence in our colony. Therefore, the range of mean feature values of highest imitation influences in our colony, but not of lowest influences, seems consistent across zebra finch colonies. This range, in turn, can be explained by balanced imitation as high influences are associated with high tutor song diversity. In sum, this outcome is consistent with the notion that over generations, songs of high feature diversity are more influential, and therefore shape the overall distribution of mean song features in a similar manner across colonies. More

1.Córdoba-Aguilar, A., González-Tokman, D. & González-Santoyo, I. Insect Behaviour (Oxford University Press, 2018).
Google Scholar 
2.Shuker, D. M. & Simmons, L. W. The Evolution of Insect Mating Systems (Oxford University Press, 2014).
Google Scholar 
3.Cade, W. Alternative male reproductive behaviors. Florida Entomol. 63, 30–35 (1980).
Google Scholar 
4.Eberhard, W. G. Copulatory courtship and cryptic female choice in insects. Biol. Rev. 66, 1–31. https://doi.org/10.1111/j.1469-185X.1991.tb01133.x (1991).
Google Scholar 
5.Matthews, R. V. & Matthews, J. R. Insect Behavior (Springer, 2010).
Google Scholar 
6.Corbet, P. S. Dragonflies: Behavior and Ecology of Odonata (Comstock Publishing Associates, 1999).
Google Scholar 
7.Córdoba-Aguilar, A. Dragonflies and Damselflies. Model Organisms for Ecological and Evolutionary Research (Oxford University Press, 2008).
Google Scholar 
8.Suhonen, J., Rantala, M. J. & Honkavaara, J. Territoriality in odonates. In Dragonflies and Damselflies: Model Organisms for Ecological and Evolutionary Research (ed. Córdoba-Aguilar, A.) 203–217 (Oxford University Press, 2008).
Google Scholar 
9.Guillermo-Ferreira, R. & Del-Claro, K. Resource defense polygyny by Hetaerina rosea Selys (Odonata: Calopterygidae): Influence of age and wing pigmentation. Neotrop. Entomol. 40, 78–84 (2011).
Google Scholar 
10.Guillermo-Ferreira, R. & Del-Claro, K. Oviposition site selection in Oxyagrion microstigma Selys, 1876 (Odonata: Coenagrionidae) is related to aquatic vegetation structure. Int. J. Odonatol. 14, 275–279 (2011).
Google Scholar 
11.Guillermo-Ferreira, R., Therézio, E. M., Gehlen, M. H., Bispo, P. C. & Marletta, A. The role of wing pigmentation, UV and fluorescence as signals in a neotropical damselfly. J. Insect Behav. 27, 67–80 (2014).
Google Scholar 
12.Guillermo-Ferreira, R., Gorb, S. N., Appel, E., Kovalev, A. & Bispo, P. C. Variable assessment of wing colouration in aerial contests of the red-winged damselfly Mnesarete pudica (Zygoptera, Calopterygidae). Sci. Nat. 102, 13 (2015).13.Guillermo-Ferreira, R. & Bispo, P. C. Male and female interactions during courtship of the Neotropical damselfly Mnesarete pudica (Odonata: Calopterygidae). Acta Ethol. 15, 173–178 (2012).
Google Scholar 
14.Gibbons, D. W. & Pain, D. The influence of river flow rate on the breeding behaviour of calopteryx damselflies. J. Anim. Ecol. 61, 283–289 (1992).
Google Scholar 
15.Robertson, H. M. Mating behaviour and its relationship to territoriality in Platycypha caligata (Selys) (Odonata: Chlorocyphidae). Behaviour 79, 11–26 (1982).
Google Scholar 
16.Conrad, K. F. & Pritchard, G. An ecological classification of odonate mating systems: The relative influence of natural, inter-and intra-sexual selection on males. Biol. J. Linn. Soc. 45, 255–269 (1992).
Google Scholar 
17.Cordero-Rivera, A. & Andrés, J. A. Male coercion and convenience polyandry in a Calopterygid damselfly (Odonata). J. Insect Sci. https://doi.org/10.1093/jis/2.1.14 (2002).
Google Scholar 
18.Miguel, T. B., Calvão, L. B., Vital, M. V. C. & Juen, L. A scientometric study of the order Odonata with special attention to Brazil. Int. J. Odonatol. 20, 27–42 (2017).
Google Scholar 
19.Berger-Tal, O. et al. A systematic survey of the integration of animal behavior into conservation. Biol. Conserv. 30, 744–753 (2016).
Google Scholar 
20.Cordero-Rivera, A. Behavioral diversity (ethodiversity): A neglected level in the study of biodiversity. Front. Ecol. Evol. 5, 1–7 (2017).
Google Scholar 
21.Guillermo-Ferreira, R. & Juen, L. Behavioral syndromes as bioindicators of anthropogenic impact. Int. J. Odonatol. (In press) (2020).22.Oliveira-Roque, F. et al. The Tinbergen shortfall: Developments on aquatic insect behavior that Are critical for freshwater conservation. In Aquatic Insects (eds Del-Claro, K. & Guillermo, R.) 365–380 (Springer, 2019).
Google Scholar 
23.Caro, T. & Sherman, P. W. Vanishing behaviors. Conserv. Lett. 5, 159–166 (2012).
Google Scholar 
24.Harabiš, F., Jakubec, P. & Hronková, J. Catch them if you can! Do traits of individual European dragonfly species affect their detectability?. Insect Conserv. Diver. https://doi.org/10.1111/icad.12378 (2019).
Google Scholar 
25.Oliveira-Junior, J. M. B. & Juen, L. The Zygoptera/Anisoptera Ratio (Insecta: Odonata): A New Tool for Habitat Alterations Assessment in Amazonian Streams. Neotrop. Entomol. 48, 552–560 (2019).
Google Scholar 
26.Silva, D. C. & Oliveira-Junior, J. M. B. Efeito da cobertura de dossel sobre a comunidade de Odonata (insecta) em igarapés na região de Santarém-Belterra (PA). Rev. Ibero-Am. Ciênc. Ambient. 9, 88–97 (2018).
Google Scholar 
27.Pereira, D. F. G., Oliveira-Junior, J. M. B. & Juen, L. Environmental changes promote larger species of Odonata (Insecta) in Amazonian streams. Ecol. Indic. 98, 179–192 (2019).
Google Scholar 
28.Carvalho, F. G., Silva-Pinto, N., Oliveira-Junior, J. M. B. & Juen, L. Effects of marginal vegetation removal on Odonata communities. Acta Limnol. Bras. 25, 10–18 (2013).
Google Scholar 
29.Rodrigues, M. E., Roque, F. O., Guillermo-Ferreira, R., Saito, V. S. & Samways, M. J. Egg-laying traits reflect shifts in dragonfly assemblages in response to different amount of tropical forest cover. Insect Conserv. Diver. 12, 231–240 (2018).
Google Scholar 
30.De Marco, P., Batista, J. D. & Cabette, H. S. R. Community assembly of adult odonates in tropical streams: An ecophysiological hypothesis. PLoS ONE 10, e0123023 (2015).
Google Scholar 
31.Seidu, I., Danquah, E., Nsor, C. A., Kwarteng, D. A. & Lancaste, L. T. Odonata community structure and patterns of land use in the Atewa Range Forest Reserve, Eastern Region (Ghana). Int. J. Odonatol. 20, 173–189 (2017).
Google Scholar 
32.Strahler, A. N. Quantitative analysis of watershed geomorphology. Eos Trans. Am. Geophys. Union. 38, 913–920 (1957).
Google Scholar 
33.Heino, J. & Peckarsky, B. L. Integrating behavioral, population and large-scale approaches for understanding stream insect communities. Curr. Opin. Insect Sci. 2, 7–13 (2014).
Google Scholar 
34.Cezário, R. R. et al. Sampling methods for dragonflies and damselflies. In Measuring Arthropod Biodiversity: A Handbook of Sampling Methods (eds Santos, J. C. & Fernandes, G. W.) 223–240 (Springer, 2020).
Google Scholar 
35.Lencioni, F. A. A. The Damselflies of Brazil in An Illustrated Guide—Coenagrionidae (All Print Editora, 2006).
Google Scholar 
36.Borror, D. J. A key to the New World genera of Libellulidae (Odonata). Ann. Entomol. Soc. Am. 38, 168–194 (1945).
Google Scholar 
37.Belle, J. Higher classification of the South-American Gomphidae (Odonata). Zool. Meded. 70, 298–324 (1996).
Google Scholar 
38.Garrison, R. W. A synopsis of the genus Hetaerina with descriptions of four new species (Odonata: Calopterigidae). Trans. Am. Entomol. Soc. 116, 175–259 (1990).
Google Scholar 
39.Lencioni, F. A. A. The Damselflies of Brazil: An Illustrated Guide—The Non Coenagrionidae Families (All Print Editora, 2005).
Google Scholar 
40.Garrison, R. W., Von Ellenrieder, N. & Louton, J. A. Dragonfly Genera of the New Word: An Illustrated and Annotated Key to the Anisoptera (The Johns Hopkins University Press, 2006).
Google Scholar 
41.Garrison, R. W., Von Ellenrieder, N. & Louton, J. A. Damselfly genera of the New World in Baltimore, An Illustrated and Annotated Key to the Zygoptera (The Johns Hopkins University Press, 2010).
Google Scholar 
42.Kaufmann, P. R., Levine, P., Robison, E. G., Seeliger, C. & Peck, D. V. Quantifying Physical Habitat in Wadeable Streams. EPA/620/R-99/003 (US Environmental Protection Agency, 1999).
Google Scholar 
43.Juen, L. et al. Effects of oil palm plantations on the habitat structure and biota of streams in Eastern Amazon. River Res. Appl. 32, 2081–2094 (2016).
Google Scholar 
44.Nessimian, J. L. et al. Land use, habitat integrity, and aquatic insect assemblages in Central Amazonian streams. Hydrobiologia 614, 117–131 (2008).
Google Scholar 
45.Brasil, L. S., Lima, E. L., Spigoloni, Z. A., Ribeiro-Brasil, D. R. G. & Juen, L. The habitat integrity index and aquatic insect communities in tropical streams: A meta-analysis. Ecol. Ind. 116, 106495 (2020).
Google Scholar 
46.Bastos, R. C. et al. Morphological and phylogenetic factors structure the distribution of damselfly and dragonfly species (Odonata) along an environmental gradient in Amazonian streams. Ecol. Indic. 122, 107257 (2021).
Google Scholar 
47.Mendoza-Penagos, C. C., Calvão, L. B. & Juen, L. A new biomonitoring method using taxonomic families as substitutes for the suborders of the Odonata (Insecta) in Amazonian streams. Ecol. Indic. 124, 107388 (2021).
Google Scholar 
48.Violle, C. et al. Let the concept of trait be functional!. Oikos 116, 882–892 (2007).
Google Scholar 
49.Luiza-Andrade, A., Assis Montag, L. F. & Juen, L. Functional diversity in studies of aquatic macroinvertebrates community. Scientometrics 111, 1643–1656 (2017).
Google Scholar 
50.Dalzochio, M. S. et al. Effect of tree plantations on the functional composition of Odonata species in the highlands of southern Brazil. Hydrobiologia 808, 283–300 (2018).
Google Scholar 
51.McCauley, S. J. Relationship between morphology, dispersal and habitat distribution in three species of Libellula (Odonata: Anisoptera). Aquat. Insects. 34, 195–204 (2012).
Google Scholar 
52.Turlure, C., Schtickzelle, N. & Baguette, M. Resource grain scales mobility and adult morphology in butterflies. Land. Ecol. 25, 95–108 (2010).
Google Scholar 
53.Reed, S. C., Williams, C. M. & Chadwick, L. E. Frequency of wing-beat as a character for separating species races and geographic varieties of Drosophila. Genetics 27, 349–361 (1942).
Google Scholar 
54.Hall, J. P. & Willmott, K. R. Patterns of feeding behaviour in adult male riodinid butterflies and their relationship to morphology and ecology. Biol. J. Linn. Soc. 69, 1–23 (2000).
Google Scholar 
55.Pavoine, S., Vallet, J., Dufour, A. B., Gachet, S. & Daniel, H. On the challenge of treating various types of variables: Application for improving the measurement of functional diversity. Oikos 118, 391–402 (2009).
Google Scholar 
56.Villéger, S., Mason, N. W. H. & Mouillot, D. New multidimensional functional diversity indices for a multifaceted framework in functional ecology. Ecology 89, 2290–2301 (2008).
Google Scholar 
57.Stine, R. A. The graphical interpretation of variance inflation factors. Am. Stat. 49, 53–56. https://doi.org/10.1080/00031305.1995.10476113 (1995).
Google Scholar 
58.Laliberté, E. & Legendre, P. A distance-based framework for measuring functional diversity from multiple traits. Ecology 91, 299–305 (2010).
Google Scholar 
59.Oksanen, J. et al. Vegan: Community Ecology Package. R package version 2.5-5, 1–298 (2019). https://CRAN.R-project.org/package=vegan.60.Chessel, D., Dufour, A. & Thioulouse, J. “The ade4 Package—I: One-Table Methods.” R News. 4, 5–10. https://cran.r-project.org/doc/Rnews/ (2004).61.Faraway, J. faraway: Functions and Datasets for Books by Julian Faraway. R package version 1, 1–117. (2016). https://CRAN.R-project.org/package=faraway62.Brasil, L. S. et al. Net primary productivity and seasonality of temperature and precipitation are predictors of the species richness of the Damselflies in the Amazon. Basic. Appl. Ecol. 35, 45–53 (2019).
Google Scholar 
63.Remsburg, A. J. & Turner, M. G. Aquatic and terrestrial drivers of dragonfly (Odonata) assemblages within and among north-temperate lakes. J. N. Am. Benthol. Soc. 28, 44–56 (2009).
Google Scholar 
64.Vilela, D. S., Ferreira, R. G. & Del-Claro, K. The Odonata community of a Brazilian vereda: Seasonal patterns, species diversity and rarity in a palm swamp environment. Bioscience 32, 486–495 (2016).
Google Scholar 
65.Sahlén, G. Specialists vs. generalists among dragonflies—The importance of forest environments to form diverse species pools. In Forests and Dragonflies (ed. Rivera, A. C.) 153–179 (Pensoft Publishers, 2006).
Google Scholar 
66.Winemiller, K. O. Ecomorphological diversification in lowland freshwater fish assemblages from five biotic regions. Ecol. Monogr. 61, 343–365 (1991).
Google Scholar 
67.Mc Peek M. A. Ecological factors limiting the distribuitions and abundances of Odonata. In Paulson, Dragonflies & Damselflies: Model Organisms for Ecological and Evolutionary Research (ed. Córdoba-Aguilar, A.) 51–62 (Oxford University Press, 2008).68.Balzan, M. V. Associations of dragonflies (Odonata) to habitat variables within the Maltese Islands: A spatio-temporal approach. J. Insect Sci. https://doi.org/10.1673/031.012.8701 (2012).
Google Scholar 
69.Ribeiro, J. R. I., Nessimian, J. L., Mendonça, E. C. & Carvalho, A. L. Aspectos da distribuição dos Nepomorpha (Hemípteros: Heterópteros) em corpos d ́água na restinga de Maricá, Estado do Rio de Janeiro. Oecol. Aust. 5, 113–128 (1998).
Google Scholar 
70.Juen, L. & De Marco, P. Dragonfly endemism in the Brazilian Amazon: Competing hypotheses for biogeographical patterns. Biodivers. Conserv. 21, 3507–3521 (2012).
Google Scholar  More

With the aim of obtaining a useful tool for stakeholders to explore, assess and use the information related to CF and WF of food commodities, we implemented a multi-step methodological framework to create an easy to use CF and WF repository of food items, which can be expanded or modified for tailored requirements using a science based approach for each step of its creation (Fig. 1).The overall methodological procedure is made of 3 steps. Step 1 is related to CF and WF data collection from literature, eligibility check and harmonization, to create the base level of the database (level 1). Step 2 is about the creation of other three informative layers with higher level of data aggregation. These might be the data of direct interests for stakeholders of the food systems. A rigorous statistical approach is proposed to evaluate the quality of analysed data and criteria for the correct use of data, based on statistical evidence, are set and applied to the data. In Step 3 the complex set of statistical evaluations, done for each informative level, is summarized into an easy to use dataset reporting values of CF and WF of food items. Thanks to its multilevel approach, the database provides a flexible tool for different purposes and levels of expertise. Each step is based on transparent procedures that allow users to replicate, to implement and to modify each level of the database.The three steps are described in details in the following paragraphs.Step 1 – CF and WF data collection, harmonization and compilation of level 1 of SEL databaseThe first step was to review the published data of CF and WF of food commodities. We revised literature data published till January 2020 including peer-reviewed papers, conference proceedings, public reports or studies where methods of data collection and handling were described, and Environmental Product Declarations (EPDs).For the collection of CF data, a significant input came from the systematic review of Clune et al.11, who reviewed 369 published studies, covering the period 2000–2015, proving 168 varieties of fresh food products based on 1718 data entries. An additional source of studies reporting both CF and WF was the Double Pyramid database 2016 built on the previous version 201414 (BCFN2016 https://www.barillacfn.com/en/publications/double-pyramid-2016/), which reports 1202 CF values from 468 sources covering 240 food items and 309 WF values from 136 data sources covering 152 food items (reference period 1998–2016). Part of CF data of this latter dataset, up to year 2014, were already revised and included in the Clune et al.11 study. To avoid double counting from these two sources, data from both sources were checked for authorship, plus the CF reported data were compared and if in disagreement the original data were checked in the paper. Data reported in the Double Pyramid database 2016 but not present in Clune et al.11, mostly referring to processed food, were checked for eligibility applying the exclusion criteria reported in Table 2 and if considered eligible they were included in the present database.Table 2 Exclusion criteria to be applied to CF and WF data collected from literature to create SEL database level 1.Full size tableA new literature search was done to integrate data not covered by the previous reviews using three online bibliographic sources SCOPUS (https://www.scopus.com/home.uri), Google Scholar (https://scholar.google.com/) and the Google search engine (https://www.google.com/), which was concluded in January 2020. To search the bibliographic sources, we used the combinations of two sets of words. The first set referred to “impacts” and included the following words: carbon footprint, water footprint, virtual water, greenhouse gases, environmental impact, life cycle, LCA, LCI, EPD. The second set referred to “products” and included words like food, beverages, fish, shellfish, crops, vegetables, fruit, meat, eggs, dairy. EPDs were updated based on data reported on the International EPD’s System database (www.environdec.com). Added studies were evaluated for exclusion criteria (Table 2).The final list of data from single studies reported in the SEL database was distributed as follow: 3349 CF data, including 1397 data of fresh food commodities already reported in Clune et al.11, 803 CF data originally reported in Double Pyramid 2016 database, which were checked for eligibility and harmonized, and 701 CF data added with this study; 938 WF data, including 288 WF data originally reported in double pyramid 2016 and 650 WF data added with this study.All the CF and WF values extracted from the collected studies were assigned a group, a typology, a sub-typology when this applied, and an item name (Table 1) and were recorded on an excel sheet including the following additional information: type of bibliographic source, full reference, publication year, system boundary at distribution, country of production, region of production, relevant notes, presence of the same value in other data collections (i.e. Clune et al.11 or Double Pyramid 2016).After data collection, CF data where further analysed and handled for the harmonization of the system boundary following the approach as reported in Clune et al.11. The system boundary considered in the SEL database is the distribution centre to consumers located in the country of origin. It hence excludes post market phase like for example cooking. The system boundaries at distribution have a wide range of specifications in the published papers. We accepted regional distribution centre (RDC), international distribution centre (IDC), European distribution centre (EDC), country ports of final destination, warehouses, wholesalers, city markets, up to retailers. For the specific case of international transport, which includes also the emissions for shipping to regional distribution centres of the hosting country, rather than excluding the studies we have created a dedicated typology “imported”, which however includes very few studies. The imported commodity is indicated in the SEL database by a capital letter “I”.If CF values collected from literature referred to the system boundary “farm gate” or “slaughterhouse”, additional post farm gate GHG emissions were added as proposed by Clune et al.11. These additional emissions also included packaging if not reported in the publication. We adopted the median value for distribution to RDC (0,09 kg CO2/kg or kg CO2/L) and packaging (0,05 kg CO2/kg or kg CO2/L) used by Clune et al.11. Data referring to slaughterhouse emissions were also taken from the same publication.To address the share of WF for packaging and transportation to the market we analysed 256 EPD’s. No significant increase of WF in downstream stages associated to packaging and distribution was found. Thus we included in the analysis all system boundaries with the exception of ‘cooking’, human excretion and waste disposal.To transform CF values from carcass or live weight to bone free meat, ratios reported in in Clune et al.11 were used, while the ratio carcass weight to bone free meat for buffalo meat (1:0.684) was estimated from the studies of Gerber et al.15, Gurunathan et al.16, Li et al.17.The final version of CF and WF data, after data handling was recorded in a sheet where, in addition to the information mentioned above for each study, we also reported additional post farm gate emissions (transport T, slaughtering S, packaging P) or meat conversion factors (cf) when applying. This complete dataset represents the level 1 information sheet of the SEL database (Fig. 1).A change in 100-year global warming potential (GWP) factors provided by the International Panel on Climate Change reports AR3 (2001), AR4 (2007) and AR5 (2013) might have introduced additional variability in the studies of LCA on which CF data of level 1 are based. The extent of such variability is difficult to quantify as it depends on the relative weight of each GHG on the total CF of the item. However, the analysis of some item groups (tomato, rice, beef meat, chicken meat), used as sample test, did not show any clear trend of CF average reduction or increase over the years (1998–2020), suggesting that differences among production processes and conditions were the dominant source of CF variability.Step 2 – Creation of derived CF and WF datasets with higher aggregation level (2, 3 and 4)This step provides footprints of food commodities with a higher level of aggregation corresponding to food items, typologies and sub-typologies (Table 1), which might be of particular interest for different kinds of stakeholders. The item represents the higher detail of aggregated footprint data of a food commodity and it is often the most desirable information for food impact analysis and dietary assessments. We propose here a methodological framework to evaluate the uncertainty associated to data used to represent food items. The methodological framework will support the users in their choice of the optimal value to represent the food item on the basis of the available data present in the database. It also would easily allow for expansion and implementation of food item values.Level 2, SEL CF ITEM & SEL WF ITEM datasetsThese two datasets (CF and WF) report a comprehensive set of descriptive statistics for the list of food items present in the database. The population of data used to attribute a value and uncertainty to a food item is made of all the CF or WF values classified with that “item entry name” in the dataset of level 1 of SEL database.The item data population is described in level 2 by the following set of information.Size: number of studies used for the analysis of item population (n).Location and central-tendency measures: in terms of mean, median, first quartile (Q1) and third quartile (Q3), including also the minimum (Min) and maximum (Max) observed values.Variability measures: Standard Deviation (SD) Coefficient of Variation (CV) as absolute and relative dispersion indexes, the Interquartile Range (IQR) and the Median Absolute Deviation (MAD) as more robust indexes of variability.Shape measures: Skewness (SK), kurtosis (KU) indexes and Shapiro-Wilk normality test (SW test).The median of the item data population was chosen to assign a value of central tendency which represents the item. The median offers the advantage of not being influenced by the presence of outliers which misrepresent the value of the mean, making it a less meaningful measure. As such, the median represents the location estimator with the highest breakdown point (equal to 0.5) and with “the maximum proportion of observations that can be contaminated (i.e., set to infinity) without forcing the estimator to result in a “false” and not-representative value18,19. With these properties, the median also represents the most appropriate measure of central tendency to describe both positively and negatively skewed distributions20.To describe the uncertainty associated to the position value (median) we used descriptive statistic data relative to dispersion and shape of item data distribution. In particular, we used skewness and kurtosis indexes, which gave us information on the existence of symmetric or skewed distributions, as well as on their ‘peakedness’ measured as relative to the weights of the tails21, thus enabling us to evaluate (for each distribution) the importance of extreme values over the entire set of data and the related level of dispersion (platykurtic versus leptokurtic distributions). We completed the shape analysis by carrying out the Shapiro-Wilk test22,23 (4 ≤ n ≤ 2000).To define the uncertainty of the item value we created an assignment method based on a combination of the three quality flags (Fig. 2).Fig. 2Method for attribution of CF (or WF) value to a food item based on data quality flags. The scheme shows the procedure applied to evaluate the level of uncertainty associated to CF or WF value of a food item and how this information is used to decide the best value that should be used to represent the item. Three quality flags related to a statistical aspect of the data population are calculated to attribute the level of uncertainty. Each flag has different level of quality, red being the worst, green the best. Flags are then combined and expert judgement is used to associate a suggestion for data use to each flag combination. If the item median value is characterized by high uncertainty it poorly represents the item and caution is needed to use this data to represent the food commodity, the users is therefore redirected to a higher level of aggregation such as the sub-typology or the typology which includes the analysed item.Full size imageFlag 1, evaluation of the ‘size’ (n) of the “item data population”

Red if n  More

Performance evaluation of tractor engineThe performance of the direct-injection turbocharger diesel engine for the Kubota M-90 tractor was evaluated at different engine loads with the use of different biodiesel blends with mineral diesel to maximize the engine efficiencies of PTO torque, BP, BMEP, and BTE, while also minimizing specific fuel consumption, gas emissions, and, finally, fossil fuel consumption. The results in Table 1 showed a significant effect of engine load percentage and fuel blend percentage and their interaction on all the studied characters.Table 1 Effects of engine load percentage and fuel blends percentage on power take-off speed, power take-off power, power take-off torque, engine speed, brake power, brake specific fuel consumption, brake thermal efficiency, fuel consumption, brake mean effective pressure, O2 percentage, CO2 percentage, CO, NO, and SO2.Full size tableEngine speedFor the effects of engine load percentage on engine speed, the results in Table 1 indicated the relationship between engine load percentage and engine speed in rpm was inversely proportional. The maximum engine speed was recorded at a loading of 0%, and the lowest speed was at a loading of 100% (Table 1). Using 100% diesel fuel (B0) gave the highest engine speed among all treatments, and the lowest speed was recorded with 100% biodiesel fuel (B100). The significant interaction between engine load, percentage, and engine speed in rpm was inversely proportional, as shown in Fig. 2a. The maximum engine speed was 2854 rpm at the loading stage of 0% using 100% diesel fuel (B0), while the minimum speed was 276 rpm at the loading stage of 100% using 100% biodiesel fuel (B100), as shown in Table 2. At all loadings, stages with increased biodiesel percentages in the blended fuel samples resulted in decreased engine speed because the heating value of biodiesel is lower than that of mineral diesel32,33,34,35.Figure 2Effects of engine load on (a) engine speed, (b) PTO torque, (c) PTO speed on PTO torque, (d) engine load on brake power, (e) engine speed on brake power, (f) engine load on fuel consumption, (g) engine speed on fuel consumption, (h) engine load on (BSFC). (i) engine speed on (BSFC), (j) engine load on BMEP, (k) engine speed on BMEP, (l) engine speed on BTE and (m) engine load on BTE.Full size imageTable 2 Interaction effects between engine load percentage and fuel blends percentage on power take-off speed, power take-off power, power take-off torque, engine speed, brake power, brake specific fuel consumption, brake thermal efficiency, fuel consumption, brake mean effective pressure, O2 percentage, CO2 percentage, CO, NO, and SO2.Full size tablePTO torqueThe results presented in Table 1 showed the significant effect of load percentage on PTO torque, where a loading stage of 75% achieved the highest PTO torque among all loading stage percentages, and the lowest value for PTO torque was obtained with a loading stage of 0%. Regarding the effects of fuel blend percentage on PTO torque, the results in Table 1 indicated that the fuel blends significantly affected PTO torque, and the highest value of this trait was achieved with B0 blend (100% diesel fuel) in comparison to the other blend percentages, while the lowest PTO torque was given with 100% biodiesel. The relationship between the torque of PTO shaft, Nm, and PTO load in percentage, and speed in rpm are shown in Fig. 2b,c, respectively. Increased PTO load resulted in decreased PTO speed and increased PTO torque until maximum torque values were reached for all blended fuel samples at a loading stage of 75% and a speed between 316 and 332 rpm, and then the torque decreased incrementally until the maximum loading stage was reached at a minimum PTO speed. Table 2 presents the results of the interactions between engine load percentage and fuel blend percentage, indicating that the maximum PTO torque was 663 Nm at a loading stage of 75% and PTO speed of 332 rpm, using 100% diesel fuel (B0), and the minimum PTO torque was 98.51 Nm at loading stage of 0% and PTO speed of 699.19, rpm using 100% diesel fuel (B0). At all loading stages, increasing biodiesel percentage in the blended fuel samples resulted in decreased PTO torque, which, due to the heating value of biodiesel, was lower than that of diesel fuel34,35,36. The values for PTO torque were close at different biodiesel percentages at the loading stage 0%, but engine performance cannot be judged at the no load stage with minimum torque, so the PTO load should be increased to see the difference between fuel types.Engine brake powerData in Table 1 showed that engine load percentage significantly affected BP, kW, such that engine load of 50% achieved the highest BP, and the lowest value for BP was obtained with 100% load. The results given in Table 1 show that fuel blend percentage was significantly affected BP, whereas the highest value for this trait was achieved with the 0% blend (100% diesel fuel) in comparison to the other blend percentage, while the lowest BP was given with 100% biodiesel. The interactions among BP, engine load, and engine speed were significant and were as presented in Fig. 2d, e. Moreover, increased engine load resulted in decreased engine speed and increasing BP until the highest value was reached at the loading stage of (50%) at engine speeds of 2034–2137 rpm for all fuel types shown in Table 2, which was due to the increased mass of burning fuel. The BP decreased until engine stop at a maximum loading stage of 100%, which was due to the effects of higher frictional force at the maximum loading stage33,34,35,37. The maximum BP was 46.2 kW at a loading stage of 50% and a speed of 2137 rpm at 100% diesel fuel (B0), while the minimum BP was 5.82 kW at a maximum loading stage of 100% and a speed of 276 rpm using 100% biodiesel (B100). At all loading stages, increased biodiesel percentages resulted in decreased BP because the calorific value of biodiesel was lower than that of diesel, as noted.Fuel consumptionData in Table 1 showed that engine load percentage affected significantly fuel consumption; 50% load achieved the highest fuel consumption, and the lowest value for fuel consumption was obtained for 0% load. The results in Table 1 showed that fuel blends percentage significantly affected fuel consumption, and the highest value of fuel consumption was recorded with the B100 blend (100% biodiesel fuel), and the lowest was given with the B0 blend of 100% diesel fuel. The significant interaction between fuel consumption, kg/h (Kilogram per hour) and each of engine load and speed are shown in Fig. 2f,g and the interaction between engine load percentage and fuel blend percentage are shown in Table 2, such that increased engine load resulted in decreased engine speed and increased fuel consumption until reaching the maximum value at a loading stage of 50% at maximum BP, which was because of the increased mass of burning fuel at this stage, and then the fuel consumption decreasing until reaching maximum loading33,34,35. The maximum fuel consumption was 18.24 kg/h at an engine speed of 2034.35 rpm using 100% biodiesel B100 at a loading stage of 50%. The minimum fuel consumption was 9.76 kg/h at an engine speed of 2692.9 rpm using 100% biodiesel B100 at a no-load stage. At loading stages between 0 and 100%, increasing biodiesel percentage resulted in increased fuel consumption, which is because the density of biodiesel was higher than that of diesel fuel.Brake specific fuel consumptionThe results in Table 1 indicated that engine load percentage significantly affected BSFC, such that the highest BSFC was achieved with an engine load of 100%, while the lowest value was obtained with an engine load of 50%. Other results shown in Table 1 indicated that increased biodiesel percentage in fuel blends produced significantly increased BSFC, and the maximum value of BSFC was given with B100 (100% biodiesel fuel); the lowest was seen with B0 percentage (100% diesel fuel). The relationship of interaction between (BSFC), (Kilogram per kilowatt hour) kg/kWh, engine load, and engine speed are shown in Fig. 2h,i, indicating that increased engine load resulted in decreased engine speed and BSFC until the minimum value was reached at a loading stage of 50% at maximum BP and fuel consumption. Then, the BSFC increased until it reached maximum value at a maximum loading stage of 100%, which was due to the highest frictional force and the lowest BP occurring at this loading stage33,34,35. The maximum (BSFC) was 1.95 kg/kWh at a loading stage of 100% and an engine speed of 276 rpm using 100% biodiesel fuel (B100); the minimum BSFC was 0.32 kg/kWh at an engine speed of 2137 rpm and a loading stage of 50% using 100% diesel fuel (B0), as shown in Table 2. At all loading stages, increased biodiesel percentages resulted in increased BSFC, except at the no loading stage. This is because the fuel consumption for biodiesel was higher than that for mineral diesel. Additionally, the calorific value of biodiesel was lower than that for diesel fuel, and the viscosity of the biodiesel was higher than that for mineral diesel, which leads to unfavorable pumping and spray characteristics36,38.Brake mean effective pressureThe results in Table 1 indicated a significant effect of engine load percentage on BMEP, such that the highest BMEP was given by an engine load of 75%, and the other side the lowest value for BMEP was obtained for an engine load of 0%. The results given in the same table indicated that increased biodiesel percentage in fuel blends significantly decreased BMEP. The maximum value of BMEP was given with the 0 blend (100% diesel fuel), and the lowest BMEP was given with 100% biodiesel fuel. The interaction between BMEP, kPa, engine load, and engine speed are shown in Fig. 2j,k. The data in Table 2 show the interaction between engine load percentage and fuel blend percentage. It can be clearly seen that increased engine load resulted in decreased engine speed and increased BMEP until the maximum value was reached at a loading stage of 75% at engine speeds between 1293 and 1355 rpm. The BMEP decreased with slight values until reaching the maximum loading stage at minimum engine speeds between 276 and 288 rpm. The maximum BMEP was 625 kPa at an engine speed of 1355 rpm, using 100% diesel fuel (B0) at a loading stage of 75%. The minimum BMEP was 92 kPa at an engine speed of 2692 rpm, using 100% biodiesel (B100) at no loading stage. At all loading stages, increased biodiesel percentage resulted in decreased BMEP, except that there was no loading stage at which the BSFC did not change with different biodiesel percentages. This is because the effect of increased engine speed resulted in a decreased time remaining for combustion and resulted in an insufficient motion of air in the cylinder. Both effects decreased the combustion efficiency and the BMEP values, as shown in Fig. 2j according to33,34,35,39.Brake thermal efficiencyThe results shown in Table 1 cleared that engine load percentage significantly affected BTE; the highest BTE was recorded with a 50% load, and the lowest one was given with a 0% load percentage. Table 1 also indicated that fuel blend percentage significantly affected BTE, and the maximum value for BTE was given with 0 blend (100% diesel fuel). The lowest value was obtained with 100% biodiesel (B100). The relationship between BTE and engine load and engine speed are shown in Figs. 2l,m. Increased engine load caused decreased engine speed and increased BTE until the maximum value was reached at a loading stage of 50%; BTE decreased until a minimum value was reached at a maximum loading stage of 100% and minimum engine speeds between 276 and 288 rpm. The maximum BTE was 26% at a speed of 2137.17 rpm using 100% diesel fuel (B0) at loading stage of 50%. The minimum BTE was 4.4% at speed of 276 rpm, using 100% biodiesel (B100) at the maximum loading stage of 100%, as shown in Table 2. For all loadings stages increased biodiesel percentage resulted in decreased BTE, except at the no loading and maximum loading stages, where the BTE did not change with different biodiesel percentages. This is because the density of waste frying oil biodiesel was higher than that of diesel fuel, while its calorific value and volatility was lower, such that the combustion characteristics of biodiesel were lower than those of diesel fuel34,35,36,40.Gas emissions qualityThe results in Table 1 showed that an engine load of 0% significantly increased O2 emissions, and fuel blends of 100% biodiesel also increased O2 emissions relative to the other treatments. The relationships between O2 emissions, biodiesel percentage, engine load, and engine speed are shown in Fig. 3a,b. Increased engine load resulted in decreased O2 emissions because of the increased engine consumption of O2 to optimize fuel combustion, while increased engine speed resulted in increased O2 emissions. The maximum O2 emissions were 15.3% at the minimum loading stage for all fuel blends, while the minimum O2 emissions were 4.3% at maximum loading stage for 100% diesel fuel, as presented in Table 2. At all loading stages, increased biodiesel percentage in the blended fuel samples resulted in increased O2 emissions, except at the no loading stage, where the oxygen content in the biodiesel was about 10 to 12% higher than that of diesel fuel34,35,41,42.Figure 3Effects of engine load on (a) engine load on O2 emissions, (b) engine speed on O2 emissions, (c) engine load on CO2 emissions, (d) engine speed on CO2 emissions, (e) engine load on CO emissions, (f) engine speed on CO emissions, (g) engine load on NO emissions, (h) engine speed on NO emissions, (i) engine load on SO2 emissions and (j) engine speed on SO2 emissions.Full size imageThe results in Table 1 showed that engine loading of 100% significantly increased CO2 and CO emissions, and the fuel blend of 100% diesel fuel (B0) increased CO2 and CO emissions relative to other treatments. The relationship between CO2 emissions, engine load, and engine speed are presented in Fig. 3c,d and Table 2. Increased engine load resulted in increased CO2 emissions until a maximum loading stage of 100% was reached, while increased engine speed resulted in decreased CO2 emissions. The maximum value for CO2 emissions was 12.3% at the maximum loading stage using 100% diesel (B0), and the minimum CO2 emissions was 4.2% at the no loading stage for all fuel blends. At loading stages of 50, 75, and 100%, increased biodiesel percentage in the blended fuel samples resulted in decreased CO2 emissions, which due to the oxygen content in the biodiesel was about 10–12%. A higher oxygen content contributes to increasing ignition quality and decrease CO2 emissions35,36,41.The relationship between CO emissions, biodiesel percentage, engine load, and engine speed are shown in Fig. 3e,f. For all tested fuel samples, increased engine load resulted in a greater increase in CO emissions, until a maximum load was reached except at 100% biodiesel (B100), which increased slightly. Increased engine speed resulted in a sharp decrease in CO emissions until the maximum speed was reached, except at B100, which decreased slightly43,44,45. The maximum CO emissions value was 369 ppm at the maximum loading stage using 100% diesel fuel (B0), while the minimum CO emissions was 69 ppm at the minimum loading stage using 100% biodiesel fuel (B100). At all loading stages, increased biodiesel percentage resulted in decreased CO emissions except that at loading stages of 25% and 50%, for which the values of CO emissions were close. This was because high oxygen content in biodiesel increases ignition quality and decreases CO emissions, so increased biodiesel percentages reduce environmental pollution36,43,44,45,46. The results in Table 2 showed that an engine load of 75% significantly increased NO emissions, and 100% biodiesel fuel (B100) increased NO emissions relative to the other treatments.The relationship between NO emissions, engine load, and engine speed are shown in Fig. 3g,h. Increased engine load resulted in decreased engine speed and increased NO emissions until the maximum value was reached at a loading stage of 75%. NO emissions decreased until reach a loading stage of 100% was reached with a minimum engine speed. The maximum NO emissions were 593 ppm at a loading stage of 75% using 100% biodiesel fuel (B100), while the minimum NO emissions were 266 ppm at the minimum loading stage using (B100) as showed in Table 1. At all loading stages, increased biodiesel percentages in the blended fuel samples resulted in increased NO emissions, except at the no loading stage, which was due to the increased burned fuel, which resulted in increased cylinder temperature. This was responsible for thermal NOx formation. Higher flame and cylinder temperatures with high oxygen content in the biodiesel led to higher NOx36,43,44,45,46. Table 2 shows that the engine load of 100% significantly increased SO2 emissions, and 100% diesel fuel increased SO2 emissions, relative to the other treatments.The relationship between SO2 emissions, diesel percentage, engine load, and engine speed are shown in Fig. 3I,j and Table 1. Increased engine load resulted in increased SO2 emissions, and increased engine speed resulted in decreased SO2 emissions. There were no SO2 emissions by using 100% biodiesel (B100). The maximum SO2 emissions was 21 ppm at maximum loading stage using 100% diesel (B0). At all loading stages increasing biodiesel percentage in the blended fuel resulted in decreasing SO2 emissions43,44,45,46. More