Philippa Kaur

Geslin, B. et al. Massively introduced managed species and their consequences for plant–pollinator interactions. Adv. Ecol. Res. 57, 147–199 (2017).
Google Scholar 
Huryn, V. M. B. Ecological impacts of introduced honey bees. Q. R. Biol. 72, 275–297 (1997).
Google Scholar 
Stout, J. C. & Morales, C. L. Ecological impacts of invasive alien species on bees. Apidologie 40, 388–409 (2009).
Google Scholar 
Hung, K.-L.J., Kingston, J. M., Albrecht, M., Holway, D. A. & Kohn, J. R. The worldwide importance of honey bees as pollinators in natural habitats. Proc. R. Soc. Ser. B 285, 20172140 (2018).
Google Scholar 
Paini, D. R. Impact of the introduced honey bee (Apis mellifera) (Hymenoptera: Apidae) on native bees: A review. Austral Ecol. 29, 399–407 (2004).
Google Scholar 
Moritz, R. F. A., Hartel, S. & Neumann, P. Global invasions of the western honey bee (Apis mellifera) and the consequences for biodiversity. Ecoscience 12, 289–301 (2005).
Google Scholar 
Paini, D. R. & Roberts, J. D. Commercial honey bees (Apis mellifera) reduce the fecundity of an Australian native bee (Hylaeus alcyoneus). Biol. Cons. 123, 103–112 (2005).
Google Scholar 
Munoz, I. & De la Rua, P. Wide genetic diversity in old world honey bees threatened by introgression. Apidologie 52, 200–217 (2021).
Google Scholar 
Williams, I. H. The dependences of crop production within the European Union on pollination by honey bees. Agric. Zool. Rev. 6, 229–257 (1994).
Google Scholar 
Thompson, C. E., Biesmeijer, J. C., Allnutt, T. R., Pietravalle, S. & Budge, G. E. Parasite pressures on feral honey bees (Apis mellifera sp.). PLoS One 9, e105164 (2014).ADS 
PubMed 
PubMed Central 

Google Scholar 
Belsky, J. & Joshi, N. K. Impact of biotic and abiotic stressors on managed and feral bees. Insects 10, 233 (2019).PubMed Central 

Google Scholar 
Medina-Flores, C. A., Guzman-Novoa, E., Hamiduzzaman, M. M., Arechiga-Flores, C. F. & Lopez-Carlos, M. A. Africanized honey bees (Apis mellifera) have low infestation levels of the mite Varroa destructor in different ecological regions in Mexico. Genet. Mol. Res. 13, 7282–7293 (2014).CAS 
PubMed 

Google Scholar 
Portman, Z. M., Tepedino, V. J., Tripodi, A. D., Szalanski, A. L. & Durham, S. L. Local extinction of a rare plant pollinator in Southern Utah (USA) associated with invasion by Africanized honey bees. Biol. Invasions 20, 593–606 (2018).
Google Scholar 
Santos, G. M. D. et al. Invasive Africanized honeybees change the structure of native pollination networks in Brazil. Biol. Invasions 14, 2369–2378 (2012).
Google Scholar 
Chapman, R. E. & Bourke, A. F. G. The influence of sociality on the conservation biology of social insects. Ecol. Lett. 4, 650–662 (2001).
Google Scholar 
Aizen, M. A. et al. When mutualism goes bad: Density-dependent impacts of introduced bees on plant reproduction. New Phytol. 204, 322–324 (2014).
Google Scholar 
Breeze, T. D. et al. Agricultural policies exacerbate honeybee pollination service supply-demand mismatches across Europe. PLoS One 9, e82996 (2014).ADS 
PubMed 
PubMed Central 

Google Scholar 
Baum, K. A. et al. Spatial distribution of Africanized honey bees in an urban landscape. Landsc. Urban Plan. 100, 153–163 (2011).
Google Scholar 
Ratnieks, F. L. W., Piery, M. A. & Cuadriello, I. The natural nest and nest density of the africanized honey-bee (Hymenoptera, Apidae) near Tapachula, Chiapas, Mexico. Can. Entomol. 123, 353–359 (1991).
Google Scholar 
Baum, K. A., Rubink, W. L., Pinto, M. A. & Coulson, R. N. Spatial and temporal distribution and nest site characteristics of feral honey bee (Hymenoptera: Apidae) colonies in a coastal prairie landscape. Environ. Entomol. 33, 727–739 (2004).
Google Scholar 
Rangel, J. et al. Africanization of a feral honey bee (Apis mellifera) population in South Texas: Does a decade make a difference?. Ecol. Evol. 6, 2158–2169 (2016).PubMed 
PubMed Central 

Google Scholar 
Oldroyd, B. P., Thexton, E. G., Lawler, S. H. & Crozier, R. H. Population demography of Australian feral bees (Apis mellifera). Oecologia 111, 381–387 (1997).ADS 
CAS 
PubMed 

Google Scholar 
Arundel, J. et al. Remarkable uniformity in the densities of feral honey bee Apis mellifera Linnaeus, 1758 (Hymenoptera: Apidae) colonies in South Eastern Australia. Austral Entomol. 53, 328–336 (2014).
Google Scholar 
Remm, J. & Lõhmus, A. Tree cavities in forests—The broad distribution pattern of a keystone structure for biodiversity. For. Ecol. Manag. 262, 579–585 (2006).
Google Scholar 
Lindenmayer, D., Crane, M., Blanchard, W., Okada, S. & Montague-Drake, R. Do nest boxes in restored woodlands promote the conservation of hollow-dependent fauna?. Restor. Ecol. 24, 244–251 (2016).
Google Scholar 
New South Wales Department of Planning, Industry and Environment 2003. https://www.environment.nsw.gov.au/topics/animals-and-plants/threatened-species/nsw-threatened-species-scientific-committee/determinations/final-determinations/2000-2003/competition-from-feral-honeybees-key-threatening-process-listing (accessed 22 Feb 2021).Goldingay, R. L., Rohweder, D. & Taylor, B. D. Nest box contentions: Are nest boxes used by the species they target?. Ecol. Manag. Restor. 21, 115–122 (2020).
Google Scholar 
Lindenmayer, D. B. et al. Are nest boxes a viable alternative source of cavities for hollow-dependent animals? Long-term monitoring of nest box occupancy, pest use and attrition. Biol. Cons. 142, 33–42 (2009).
Google Scholar 
Lindenmayer, D. B. et al. The anatomy of a failed offset. Biol. Conserv. 210, 286–292 (2017).
Google Scholar 
Macak, P. V. Nest boxes for wildlife in Victoria: An overview of nest box distribution and use. Vic. Nat. 137, 4–14 (2020).
Google Scholar 
Le Roux, D. S. et al. Effects of entrance size, tree size and landscape context on nest box occupancy: Considerations for management and biodiversity offsets. For. Ecol. Manag. 366, 135–142 (2016).
Google Scholar 
Berris, K. K. & Barth, M. PVC nest boxes are less at risk of occupancy by feral honey bees than timber nest boxes and natural hollows. Ecol. Manag. Restor. 21, 155–157 (2020).
Google Scholar 
Jaffe, R. et al. Estimating the density of honeybee colonies across their natural range to fill the gap in pollinator decline censuses. Conserv. Biol. 24, 583–593 (2010).PubMed 

Google Scholar 
Utaipanon, P., Schaerf, T. M. & Oldroyd, B. P. Assessing the density of honey bee colonies at ecosystem scales. Ecol. Entomol. 44, 291–304 (2019).
Google Scholar 
Utaipanon, P., Holmes, M. J., Chapman, N. C. & Oldroyd, B. P. Estimating the density of honey bee (Apis mellifera) colonies using trapped drones: Area sampled and drone mating flight distance. Apidologie 50, 578–592 (2019).CAS 

Google Scholar 
Williamson, E. M. Reliability of honey bee hive density estimates using drone sampling: does relative hive size or distance affect a colony’s drone contribution? Honours Thesis, The University of Adelaide (2020).Benson, J. S. The effect of 200 years of European settlement on the vegetation and flora of New South Wales. Cunninghamia 2, 343–370 (1991).
Google Scholar 
New South Wales Office of Environment and Heritage 2015. Upgraded NSW woody vegetation extent for 2011. http://data.auscover.org.au/xwiki/bin/view/Product+pages/nsw+5m+woody+extent+and+fpc (accessed 13 May 2020).R Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, 2020). www.R-project.org (accessed 12 January 2021).Burnham, K. P. & Anderson, D. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach (Springer, 2002).MATH 

Google Scholar 
Albert, A. & Anderson, J. A. On the existence of maximum likelihood estimates in logistic regression models. Biometrika 71, 1–10 (1984).MathSciNet 
MATH 

Google Scholar 
Firth, D. Bias reduction of maximum likelihood estimates. Biometrika 80, 27–38 (1993).MathSciNet 
MATH 

Google Scholar 
Kosmidis, I., Pagui, E. C. K. & Sartori, N. Mean and median bias reduction in generalized linear models. Stat. Comput. 30, 43–59 (2020).MathSciNet 
MATH 

Google Scholar 
Anderson, D. R. Model Based Inference in the Life Sciences: A Primer on Evidence (Springer Science & Business Media, 2007).
Google Scholar 
Barton, K. MuMIn: Multi-model inference. R package version 1.43.17 (2016).Hijmans, R. J. Raster: Geographic Data Analysis and Modeling. R package version 3.4-5 (2020).Kassambara, A. ggpubr: ‘ggplot2’ Based Publication Ready Plots. R package version 0.3.0 (2018).Kosmidis, I. brglm2: Bias Reduction in Generalized Linear Models. R package version 0.6.2 (2020).Kosmidis, I., Schumacher, D. detectseparation: Detect and Check for Separation and Infinite Maximum Likelihood Estimates. R package version 0.1 (2020).Pebesma, E. Simple features for R: Standardized support for spatial vector data. R J. 10, 439–446 (2018).
Google Scholar 
Pateiro-Lopez, B., Rodriguez-Casal, A. Alphahull: Generalization of the Convex Hull of a Sample of Points in the Plane. R package version 2.2 (2019).Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer, 2009).MATH 

Google Scholar 
Wickham, H. The split-apply-combine strategy for data analysis. J. Stat. Softw. 40, 1–29 (2011).
Google Scholar 
Wickham, H. Forcats: Tools for working with categorical variables (factors). R package version 0.5.0 (2018).Wickham, H., François, R., Henry, L., Müller, K. dplyr: A Grammar of Data Manipulation. R package version 1.0.0 (2021).Birtchnell, M. J. & Gibson, M. Long-term flowering patterns of melliferous Eucalyptus (Myrtaceae) species. Aust. J. Bot. 54, 745–754 (2006).
Google Scholar 
Steinhauer, N. et al. Drivers of colony losses. Curr. Opin. Insect Sci. 26, 142–148 (2018).PubMed 

Google Scholar 
Cunningham, S. A., Heard, T. & FitzGibbon, F. The future of pollinators for Australian Agriculture. Aust. J. Agric. Res. 53, 893–900 (2002).
Google Scholar 
Hinson, E. M., Duncan, M., Lim, J., Arundel, J. & Oldroyd, B. P. The density of feral honey bee (Apis mellifera) colonies in South East Australia is greater in undisturbed than in disturbed habitats. Apidologie 46, 403–413 (2015).
Google Scholar 
McIntyre, S. Ecological and anthropomorphic factors permitting low-risk assisted colonization in temperate grassy woodlands. Biol. Conserv. 144, 1781–1789 (2011).
Google Scholar 
Steffan-Dewenter, I. & Kuhn, A. Honeybee foraging in differentially structured landscapes. Proc. R. Soc. B Biol. Sci. 270, 569–575 (2003).
Google Scholar 
Wintle, B. A. et al. Global synthesis of conservation studies reveals the importance of small habitat patches for biodiversity. Proc. Natl. Acad. Sci. U.S.A. 116, 909–914 (2019).CAS 
PubMed 

Google Scholar 
Arthur, A. D., Li, J., Henry, S. & Cunningham, S. A. Influence of woody vegetation on pollinator densities in oilseed Brassica fields in an Australian temperate landscape. Basic Appl. Ecol. 11, 406–414 (2010).
Google Scholar 
Lindenmayer, D. B. et al. New policies for old trees: Averting a global crisis in a keystone ecological structure. Conserv. Lett. 7, 61–69 (2014).
Google Scholar 
Crane, M. J., Lindenmayer, D. B. & Cunningham, R. B. The value of countryside elements in the conservation of a threatened arboreal marsupial Petaurus norfolcensis in agricultural landscapes of south-eastern Australia—the disproportional value of scattered trees. PLoS One 9, e107178 (2014).ADS 
PubMed 
PubMed Central 

Google Scholar 
Gibbons, P., Lindenmayer, D. B., Barry, S. C. & Tanton, M. T. Hollow selection by vertebrate fauna in forests of southeastern Australia and implications for forest management. Biol. Conserv. 103, 1–12 (2002).
Google Scholar 
Seeley, T. D. & Morse, R. A. The nest of the honey bee (Apis mellifera L.). Insectes Soc. 23, 495–512 (1976).
Google Scholar 
Hung, K. L. J., Ascher, J. S., Davids, J. A. & Holway, D. A. Ecological filtering in scrub fragments restructures the taxonomic and functional composition of native bee assemblages. Ecology 100, e02654 (2019).PubMed 

Google Scholar 
Cockle, K. L., Martin, K. & Drever, M. C. Supply of tree-holes limits nest density of cavity-nesting birds in primary and logged subtropical Atlantic forest. Biol. Conserv. 143, 2851–2857 (2010).
Google Scholar 
Heard, T. Stingless bees. In Australian Native Bees: A Practical Hand Book 106–139 (NSW Department of Primary Industries, 2016).Geoscience Australia 2006. GEODATA TOPO 250K. Commonwealth of Australia. http://pid.geoscience.gov.au/dataset/ga/63999 (accessed 11 December 2020). More

Esperanza Martínez-Romero is a professor of ecological genomics and was coordinator of the undergraduate programme on genomics at Universidad Nacional Autónoma de México. Her work on plant symbioses, and outreach with local farmers has encouraged uptake of sustainable practices and the use of biofertilizers.It was during my first year as an undergraduate student that I was exposed to genetic engineering, when Dr Francisco Bolívar lectured on his development of vectors for gene cloning. I found these results fascinating, and it was listening to talks from scientists at my institute that made me realize that research was my vocation. Towards the end of my bachelor’s degree, Dr Marc von Montagu from Belgium visited and told us about plant genetic transformations — a new field within genetic engineering. Although I was accepted into his laboratory to do my doctorate, I preferred Mexico. I turned my academic journey around and instead chose to apply to a new research centre in Cuernavaca outside of Mexico City — my next turning point. I suspected that a new research centre would provide more opportunities for the development of novel areas, and would have open positions for researchers. Indeed, I was hired at this new research centre and started my own ecology group. It was there that I started working with nitrogen-fixing bacteria and plants. The effects of nitrogen-fixing bacteria on plants were outstanding. Although the scope of molecular biology was incipient to the characterization of bacterial species and populations, we were nevertheless able to make molecular characterizations of the rhizobial species that formed nitrogen-fixing nodules on beans — the most important legume for human consumption in the world. In 1991, we described a novel species, Rhizobium tropici, which could deliver high levels of nitrogen to legumes. It was then that I realized nitrogen fixation is key to the development of sustainable agriculture and could benefit farmers in Mexico and around the world. Some of the species described by my group are now used as inoculants in agriculture, reducing the use of chemical fertilizers and allowing farmers to make cost savings. To facilitate this, I published a manual on biofertilization for farmers and gave conferences and workshops to them. My group has also undertaken reforestation programmes using nitrogen-fixing legume trees inoculated with the rhizobial species that we described. More

Cape gannet movement trackingThe study took place in the Western Cape, South Africa, where we studied chick-rearing Cape gannets from Malgas Island (33.05° S, 17.93° E) during October–November from 2008 to 2015 (Fig. 1). We caught birds using a pole fitted with a loop and fitted 197 adult Cape gannets (22 in 2008, 16 in 2009, 38 in 2010, 11 in 2011, 29 in 2012, 29 in 2013, 23 in 2014, and 29 in 2015) with GPS-loggers (2008: GPS mass 65 g, i.e., 2.4 % adult body mass, Technosmart, Rom. 2009–2010: GPS mass 45 g, i.e., 1.7% adult body mass, Technosmart, Rom. From 2011: GPS mass 30 g, 1.1% of bird body mass, Catnip Technologies, Hong-Kong). Loggers were attached to the lower back with waterproof Tesa® tape and recorded position at a regular 30-s to 2-min intervals, reinterpolated over 1-min intervals. Devices were recovered after one foraging trip lasting a few hours to one week. Bird handling and tracking using these procedures do not have a measurable impact on foraging behavior19,20. We caught adult birds at-random from the colony, and previous studies showed that this resulted in a well-balanced sex-ratio preventing confounding sex effects21. All experiments were performed under permit from South African National Parks with respect to animal ethics (N° RYAP/AGR/001-2002/V1).Cape gannet movement tactics and behavioral phasesWe identified two movement trip tactics for Cape gannets: After their daytime foraging activities, some birds returned to the colony at night (rest at colony tactic) while others spent all the night at sea (rest at sea tactic). Within the GPS tracks of gannets from these two categories, we discriminated resting, foraging, and commuting phases, with a segmentation-clustering method based on smoothed speed (i.e., speed smoothed over two steps before and after the focal location) and turning angle measured at constant step length. This corresponded to the angle between the focal location, the first location entering a circle of radius equal to the median step length, and the last location inside the circle22. We fitted behavioral identification with the segclust2d package23 for the R software24. See complete details on behavioral classification for Cape gannets tracks in Appendix 1 in Courbin et al.25.Cape fur seal movement tracking and the seascape of fearWe assessed the at-sea spatial distribution of Cape fur seals, a predator of Cape gannet fledglings7 and adults (Supplementary Data 1). We used Argos data collected from 25 lactating female seals before (2003 and 2004) and again concomitantly with gannet tracking (2012 and 2014). Seals were tracked during the same period of the year as gannets (i.e., September to November). Adult females nursing pups were selected at random and captured using a modified hoop net. Once restrained, anesthesia was induced using isoflurane gas delivered via a portable vaporizer (Stinger, Advanced Anesthesia Specialists, Gladesville, New South Wales, Australia). A satellite tag was glued to the guard hairs on the upper back. Individuals were allowed to recover from the anesthesia and resumed normal behavior within 45 min of capture. Throughout the process, the animals’ breathing was closely monitored and their flippers were repeatedly flushed with seawater to prevent hyperthermia. Seals were equipped with Argos satellite transmitters at three colonies (Fig. 1): Kleinsee (29°35’09”S, 16°59’56”E) located ~400 km to the North of the gannet colony (n = 8 seals in 2003 and 2004); Vondeling Island (33°09’11”S, 17°58’57”E), ~12 km away from the gannet colony (n = 12 seals in 2012 and 2014); and Geyser Rock (34°41’19”S, 19°24’49”E) located ~230 km to the South of the gannet colony (n = 5 seals in 2003). Seals at Vondeling Island were equipped with Argos-linked Spot-6 position transmitting tags (Wildlife Computers) following deployment procedures outlined in Kirkman et al.26. Seals at Kleinsee and Geyser Rock were equipped with ST18 and ST20 satellite-linked platform terminal transmitters (Telonics, Mesa, USA), as detailed in Skern-Mauritzen et al.27. Devices collected a well-balanced number of Argos locations during the day (n = 6080 locations) and at night (n = 6501 locations). See full details on seal tracking in Supplementary Table 6. All fieldwork was permitted by the Animal Ethics Committee of the Department of Environmental Affairs and Tourism’s Marine and Coastal Management branch, which at the time was the management authority of South Africa’s marine and coastal environment (Ref: DEAT2006-06-23).We modeled both daytime and nighttime at-sea occurrences of seals for each colony with resource selection functions (RSF)28,29, a proxy of the fear effect for Cape gannets. RSF compared environmental features of seal’s at-sea Argos positions (i.e., further 500 m than the colony) with five times more random locations that captured the breadth of environmental conditions available to seals. We sampled random locations for each individual within the yearly area used by seals from each colony, delineated by the 95% kernel utilization distribution of the Argos locations of all seals of the colony. RSF were fitted with a generalized linear mixed model with a binomial distribution for errors. As environmental variables, we considered bathymetry (m), the slope of the bathymetry (°) and the distance to the colony (km) within the RSF. These variables were not highly correlated (|r| ≤ 0.61) and had low collinearity with a variance inflation factor VIF  More

Alberdi, M. T. A review of Old World hipparionine horses in The Evolution of Perissodactyls (eds. Prothero, D. R. & Schoch, R. M.) 234–261 (Clarendon Press, Oxford University Press, New York, NY ⋅ Oxford, 1989).Sen S. Hipparion Datum and its chronologic evidence in the Mediterranean area in European Neogene Mammal Chronology (eds. Lindsay, E. H., Fahlbusch, V. & Mein, P.) 495–506 (Plenum Press, New York, 1990).Garcés, M., Cabrera, L., Agustí, J. & Parés, J. M. Old World first appearance datum of “Hipparion” horses: Late Miocene large–mammal dispersal and global events. Geology 25(1), 19–22. https://doi.org/10.1130/0091-7613(1997)025%3c0019:OWFADO%3e2.3.CO;2 (1997).ADS 
Article 

Google Scholar 
Woodburne, M. O. A new occurrence of Cormohipparion, with implications for the Old World Hippotherium Datum. J. Vert. Paleont. 25(1), 256–257 (2005).Woodburne, M. O. Phyletic diversification of the Cormohipparion occidentale complex (Mammalia; Perissodactyla, Equidae), Late Miocene, North America, and the origin of the Old World Hippotherium Datum. B. Am. Mus. Nat. Hist. 306, 1–180 (2007).Article 

Google Scholar 
Bernor, R. L., Qiu, Z. & Tobien, H., 1987. Phylogenetic and biogeographic bases for an Old World hipparionine horse geochronology. Proceedings of the VIIIth International Congress of the Regional Committee on Mediterranean Neogene Stratigraphy, Budapest. Ann. Inst. Geol. Publ. Hung. 70, 43–53 (1987).Bernor, R. L., Tobien, H., Hayek, L–A. C. & Mittmann, H. –W. Hippotherium primigenium (Equidae, Mammalia) from the late Miocene of Höwenegg (Hegau, Germany). Andrias 10, 1–230 (1997).Qiu, Z., Huang, W. & Guo, Z. The Chinese hipparionine fossils. Palaeont. Sin. New Ser C 25, 1–250 (1987) ((in Chinese with English summary)).
Google Scholar 
Zouhri, S. & Bensalmia, A. Révision systématique des Hipparion sensu lato (Perissodactyla, Equidae) de l’Ancien Monde. Estud. Geol. 61, 61–99. https://doi.org/10.3989/egeol.05611-243 (2005).Article 

Google Scholar 
Liu, Y. Late Miocene hipparionine fossils from Lantian, Shaanxi Province and phylogenetic analysis on Chinese Hipparionines. PhD thesis (University of Chinese Academy of Sciences, Beijing) (2013).Bernor, R. L., Wang, S., Liu, Y., Chen, Y. & Sun, B. Shanxihippus dermatorhinus (new gen.) with comparisons to Old World hipparions with specialized nasal apparati. Riv. Ital. Paleontol. Stratigr. 124, 361–386 (2018).Bernor, R. L., Boaz, N. T., Omar, C., El-Shawaihdi, M. H. & Rook, L. Sahabi Eurygnathohippus feibeli: Its systematic, stratigraphic, chronologic and biogeographic contexts. Riv. Ital. Paleontol. Stratigr. 126, 561–581 (2020).
Google Scholar 
Liu, T., Li, C. & Zhai, R. Pliocene mammalian fauna of Lantian, Shaangxi. Prof. Pap. Stratigr. Paleont. 7, 149–200 (1978) (in Chinese).Deng, T. et al. Locomotive implication of a Pliocene three–toed horse skeleton from Tibet and its paleo–altimetry significance. Proc. Natl. Acad. Sci. USA 109, 7374–7378. https://doi.org/10.1073/pnas.1201052109 (2012).ADS 
Article 
PubMed 
PubMed Central 

Google Scholar 
Fang, X., Garzione, C., van der Voo, R., Li, J. & Fan, M. Flexural subsidence by 29 Ma on the NE edge of Tibet from the magnetostratigraphy of Linxia Basin China. Earth Planet. Sci. Lett. 210, 545–560 (2003).ADS 
CAS 
Article 

Google Scholar 
Fang, X. et al. Tectonosedimentary evolution model of an intracontinental flexural (foreland) basin for paleoclimatic research. Glob. Planet Change 145, 78–97. https://doi.org/10.1016/j.gloplacha.2016.08.015 (2016).ADS 
Article 

Google Scholar 
Deng, T., Qiu, Z., Wang, B., Wang, X. & Hou, S. Chapter 9: Late Cenozoic Biostratigraphy of the Linxia Basin, Northwestern China in Fossil Mammals of Asia: Neogene Biostratigraphy and Chronology (eds. Wang, X., Flynn, L. J. & Fortelius, M.) 243–273 (Columbia University Press, New York, 2013).Li, Y., Deng, T., Hua, H., Li, Y. & Zhang, Y. Assessment of dental ontogeny in late Miocene hipparionines from the Lamagou fauna of Fugu, Shaanxi Province China. PLoS ONE https://doi.org/10.1371/journal.pone.0175460 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Li, Y., Deng, T., Hua, H., Sun, B. & Zhang, Y. Locomotor adaptations of 7.4 Ma Hipparionine fossils from the middle reaches of the Yellow River and their palaeoecological significance. Hist. Biol. 33(7), 927–940 (2021).Bernor, R. L. & Sun, B. Morphology through ontogeny of Chinese Proboscidipparion and Plesiohipparion and observations on their Eurasian and African relatives. Vert. PalAsiat. 53, 73–92 (2015).
Google Scholar 
Woodburne, M. O. & Bernor, R. L. On superspecific groups of some Old World hipparinonine horses. J. Paleontol. 54(6), 1319–1348 (1980).
Google Scholar 
Bernor, R. L., Kaya, F., Kaakinen, A., Saarinen, J. & Fortelius, M. Old World hipparion evolution, biogeography, climatology and ecology. Earth Sci. Rev. 211, 103784 (2021).Article 

Google Scholar 
Bernor, R. L., Göhlich, U. B., Harzhauser, M. & Semprebon, G. M. The Pannonian C hipparions from the Vienna Basin. Palaegeogr. Palaeoclim. Palaeoecol. 476, 28–41. https://doi.org/10.1016/j.palaeo.2017.03.026 (2017).ADS 
Article 

Google Scholar 
Arambourg, C. Vertebres continentaux du Miocene superieur de l’Afrique du Nord. Service Carte Geologie Algerie Paleontologie Memoire, Nouveaux Serie 4, 1–159 (1959).
Google Scholar 
Bernor, R.L. & White, T.D. Systematics and biogeography of “Cormohipparion” africanum, Early Vallesian (MN 9, ca. 10.5 Ma) of Bou Hanifia, Algeria in Papers on Geology, Vertebrate Paleontology, and Biostratigraphy in Honorof Michael O. Woodburne. (ed. Albright, B.), Bull., Mus. of No. Arizona. 65, 635–658 (2009).Bernor, R. L., Scott, R. S., Fortelius, M., Kappelman, J. & Sen, S. Systematics and Evolution of the late Miocene Hipparions from Sinap, Turkey in The Geology and Paleontology of the Miocene Sinap Formation, Turkey (eds. Fortelius, M., Kappelman, J., Sen, S. & Bernor, R. L.) 220–281 (Columbia University Press, New York, 2003).Sack, W. O. The stay–apparatus of the horse’s hindlimb, explained. Equine Pract. 11, 31–35 (1988).
Google Scholar 
MacFadden, B. J. In Fossil Horses: Systematics, Paleobiology, and Evolution of the Family Equidae (Cambridge University Press, 1992).
Google Scholar 
Li, F. & Li D. Latest Miocene Hipparion (Plesiohipparion) of Zanda Basin in Paleontology of the Ngari Area, Tibet (Xizang) (eds. Yang, Z. & Nie, Z.) 186–193 (China University of Geosciences Press, Wuhan, 1990).Thomason, J. J. The functional morphology of the manus in tridactyl equids Merychippus and Mesohippus: Paleontological inferences from neontological models. J. Vert. Paleont. 6, 143–161 (1986).Article 

Google Scholar 
Eisenmann, V. What metapodial morphometry has to say about some Miocene Hipparions in Paleoclimate and Evolution, with Emphasis on Human Origins (eds. Vrba, E. S., Denton, G. H., Partridge, T. C. & Burckle, L. H.) 148–164 (Yale University Press, New Haven, 1995).Deng, T. & Wang, X. Late Miocene Hipparion (Equidae, Mammalia) of eastern Qaidam Basin in Qinghai, China. Vert. PalAsiat. 42(4), 316–333 (2004) (in Chinese with English summary).Wang, X. et al. Vertebrate paleontology, biostratigraphy, geochronology, and paleoenvironment of Qaidam Basin in northern Tibetan Plateau. Palaegeogr. Palaeoclim. Palaeoecol. 254, 363–385 (2007).Article 

Google Scholar 
Zhang, Z. et al. Chapter 6: Mammalian Biochronology of the Late Miocene Bahe Formation in Fossil Mammals of Asia: Neogene Biostratigraphy and Chronology (eds. Wang, X., Flynn, L. J. & Fortelius, M.) 187–202 (Columbia University Press, New York, 2013).Xue, X. X., Zhang, Y. X. & Yue, L. P. Discovery of Hipparion fauna of Laogaochuan and its division of eras, Fugu County Shaanxi. Chin. Sci. Bull. 40, 447–449 (1995).Article 

Google Scholar 
Deng, T. Late Cenozoic environmental changes in the Linxia Basin (Gansu, China) as indicated by cenograms of fossil mammals. Vert. PalAsiat. 47(4), 282–298 (2009).
Google Scholar 
An, Z., Kutzbach, J. E., Prell, W. L. & Porter, S. C. Evolution of Asian monsoons and phased uplift of the Himalaya-Tibetan plateau since Late Miocene times. Nature 411, 62–66. https://doi.org/10.1038/35075035 (2001).ADS 
CAS 
Article 

Google Scholar 
An, Z. et al. Changes of the monsoon–arid environment in China and growth of the Tibetan Plateau since the Miocene. Q. Sci. 26(5), 678–693. https://doi.org/10.3321/j.issn:1001-7410.2006.05.002 (2006).Article 

Google Scholar 
Wang, X. et al. Origin of the Red Earth sequence on the northeastern Tibetan Plateau and its implications for regional aridity since the middle Miocene. Sci. China D Earth Sci. 49(5), 505–517. https://doi.org/10.1007/s11430-006-0505-3 (2006).ADS 
CAS 
Article 

Google Scholar 
Dettman, D. L., Fang, X., Garzione, C. N. & Li, J. Uplift–driven climate change at 12 Ma: A long δ18O record from the NE margin of the Tibetan plateau. Earth Planet. Sci. Lett. 214, 267–277. https://doi.org/10.1016/S0012-821X(03)00383-2 (2003).ADS 
CAS 
Article 

Google Scholar 
Jiang, H. C. & Ding, Z. L. A 20 Ma pollen record of East-Asian summer monsoon evolution from Guyuan, Ningxia China. Palaegeogr. Palaeoclim. Palaeoecol. 265, 30–38. https://doi.org/10.1016/j.palaeo.2008.04.016 (2008).ADS 
Article 

Google Scholar 
Fortelius, M. et al. Late Miocene and Pliocene large land mammals and climatic changes in Eurasia. Palaegeogr. Palaeoclim. Palaeoecol. 238, 219–227. https://doi.org/10.1016/j.palaeo.2006.03.042 (2006).ADS 
Article 

Google Scholar 
Fortelius, M. et al. Evolution of neogene mammals in Eurasia: Environmental forcing and biotic interactions. Annu. Rev. Earth Pl Sci. 42, 579–604 (2014).ADS 
CAS 
Article 

Google Scholar 
Bernor, R. L., Scott, R. S., Fortelius, M., Kappelman, J. & Sen, S. Equidae (Perissodactyla) in The Geology and Paleontology of the Miocene Sinap Formation, Turkey. (eds. Fortelius, M., Kappelman, J., Sen, S. & Bernor, R.L.) 220–281 (Columbia University Press, New York, 2003).Zhang, Z. S., Wang, H. J., Guo, Z. T. & Jiang, D. B. What triggers the transition of palaeoenvironmental patterns in China, the Tibetan Plateau uplift or the Paratethys Sea retreat?. Palaegeogr. Palaeoclim. Palaeoecol. 245, 317–331. https://doi.org/10.1016/j.palaeo.2006.08.003 (2007).ADS 
Article 

Google Scholar 
Böhme, M., Ilg, A. & Winklhofer, M. Late Miocene “washhouse” climate in Europe. Earth Planet. Sci. Lett. 275, 393–401. https://doi.org/10.1016/j.epsl.2008.09.011 (2008).ADS 
CAS 
Article 

Google Scholar 
Janis, C. M., Damuth, J. & Theodor, J. M. Miocene ungulates and terrestrial primary productivity: Where have all the browsers gone?. Proc. Natl. Acad. Sci. USA 97(14), 7899–7904. https://doi.org/10.1073/pnas.97.14.7899 (2000).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Janis, C. M., Damuth, J. & Theodor, J. M. The origins and evolution of the North American grassland biome: The story from the hoofed mammals. Palaegeogr. Palaeoclim. Palaeoecol. 177, 183–198. https://doi.org/10.1016/S0031-0182(01)00359-5 (2002).ADS 
Article 

Google Scholar 
Mihlbachler, M. C., Rivals, F., Solounias, N. & Semprebon, G. M. Dietary change and evolution of horses in North America. Science 331, 1178–1181. https://doi.org/10.1126/science.1196166 (2011).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Hayek, L. C., Bernor, R. L., Solounias, N. & Steigerwald, A. Preliminary studies of hipparionine horse diet as measured by tooth microwear. Ann. Zool. Fenniuci 28, 187–200 (1992).
Google Scholar 
Sisson, S. The anatomy of the domestic animals (Saunders W B Comp, 1953).
Google Scholar 
Budras, K.-D., Sack, W. O. & Röck, S. In Anatomy of the horse. (Schlütersche, Hannover, 2009).Eisenmann, V., Alberdi, M. T., de Giuli, C. & Staesche, U. In Studying Fossil Horses, Vol. I: Methodology. (ed. Brill, E. J.) 1–71 (Leiden, 1988).Deng, T., Hou, S. & Wang, S. Neogene integrative stratigraphy and timescale of China. Sci. China Earth Sci. 62, 310–323 (2019).ADS 
CAS 
Article 

Google Scholar  More

Witman, J. D. & Smit, F. Rapid community change at a tropical upwelling site in the Galapagos Marine Reserve. Biodivers. Conserv. 12, 25–45 (2003).
Google Scholar 
Edgar, G. J., Banks, S., Fariña, J. M., Calvopiña, M. & Martínez, C. Regional biogeography of shallow reef fish and macro-invertebrate communities in the Galapagos archipelago. J. Biogeogr. 31, 1107–1124 (2004).
Google Scholar 
Okey, T. A. et al. A trophic model of a Galápagos subtidal rocky reef for evaluating fisheries and conservation strategies. Ecol. Model. 172, 383–401 (2004).
Google Scholar 
Briggs, J. C. & Bowen, B. W. A realignment of marine biogeographic provinces with particular reference to fish distributions. J. Biogeogr. 39, 12–30 (2012).
Google Scholar 
Salinas de León, P. et al. Largest global shark biomass found in the northern Galápagos Islands of Darwin and Wolf. PeerJ 4, e1911. https://doi.org/10.7717/peerj.1911 (2016).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
Humann, P. & DeLoach, N. Reef Fish Identification: Galápagos (ed. Humann, P.) (New World Publications, Inc., 2003).McCosker, J. E. & Rosenblatt, R. H. The fishes of the Galápagos archipelago: An update. Proc. Calif. Acad. Sci. 61, 167–195 (2010).
Google Scholar 
Grove, J. S. & Lavenberg, R. J. The Fishes of the Galapagos Islands (Stanford University Press, 1997).Allen, G. & Ross-Robertson, D. Fishes of Tropical Eastern Pacific (University of Hawaii Press, 1994).Ruttenberg, B. I., Haupt, A. J., Chiriboga, A. I. & Warner, R. R. Patterns, causes and consequences of regional variation in the ecology and life history of a reef fish. Oecologia 145, 394–403 (2005).ADS 
PubMed 

Google Scholar 
Bernardi, G. et al. Darwin’s fishes: Phylogeography of Galápagos Islands reef fishes. Bull. Mar. Sci. 90, 533–549 (2014).
Google Scholar 
Banks, S., Vera, M. & Chiriboga, A. Establishing reference points to assess long-term change in zooxanthellate coral communities of the northern Galápagos coral reefs. Galapagos Res. 66, 43–64 (2009).
Google Scholar 
Palacios, D., Bograd, S., Foley, D. & Schwing, F. Oceanographic characteristics of biological hot spots in the North Pacific: A remote sensing perspective. Deep Sea Res Part II Top. Stud. Oceanogr. 53, 250–269 (2006).ADS 

Google Scholar 
Sweet, W. V. et al. Water mass seasonal variability in the Galapagos Archipelago. Deep Sea Res. Part I Oceanogr. Res. Pap. 54, 2023–2035 (2007).ADS 

Google Scholar 
Schaeffer, B. et al. Phytoplankton biomass distribution and identification of productive habitats within the Galapagos Marine Reserve by MODIS, a surface acquisition system, and in-situ measurements. Remote Sens. Environ. 112, 3044–3054 (2008).ADS 

Google Scholar 
Witman, J. D., Brandt, M. & Smith, F. Coupling between subtidal prey and consumers along a mesoscale upwelling gradient in the Galapagos Islands. Ecol. Monogr. 80, 153–177 (2010).
Google Scholar 
Moity, N. Evaluation of no-take zones in the Galápagos marine reserve, zoning plan 2000. Frontiers. 5, 244. https://doi.org/10.3389/fmars.2018.00244 (2018).Article 

Google Scholar 
Lamb, R. W., Smith, F. & Witman, J. D. Consumer mobility predicts impacts of herbivory across an environmental stress gradient. Ecology 101, e02910. https://doi.org/10.1002/ecy.2910 (2020).Article 
PubMed 

Google Scholar 
Edgar, G. J. et al. Conservation of threatened species in the Galapagos Marine Reserve through identification and protection of marine key biodiversity areas. Aquat. Conserv. 18, 955–968 (2008).
Google Scholar 
Carrión-Cortez, J. A., Zárate, P. & Seminoff, J. A. Feeding ecology of the green sea turtle (Chelonia mydas) in the Galapagos Islands. J. Mar. Biol. Assoc. U. K. 90, 1005–1013 (2010).
Google Scholar 
Moity, N., Delgado, B. & Salinas-de-León, P. Correction: Mangroves in the Galapagos islands: Distribution and dynamics. PLoS One 14, e0212440. https://doi.org/10.1371/journal.pone.0212440 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Seitz, R. D., Wennhage, H., Bergström, U., Lipcius, R. N. & Ysebaert, T. Ecological value of coastal habitats for commercially and ecologically important species. ICES J. Mar. Sci. 71, 648–665 (2014).
Google Scholar 
Aguaiza, C. The role of mangrove as nursery habitats for coral reef fish species in the Galapagos Islands. MSc Thesis (University of Queensland, 2016).Llerena-Martillo, Y., Peñaherrera-Palma, C. & Espinoza, E. Fish assemblages in three fringed mangrove bays of Santa Cruz Island, Galapagos Marine Reserve. Rev. Biol. Trop. 66, 674–687 (2018).
Google Scholar 
Fierro-Arcos, D. et al. Mangrove fish assemblages reflect the environmental diversity of the Galapagos Islands. Mar. Ecol. Prog. Ser. 664, 183–205 (2021).ADS 

Google Scholar 
Henseler, C. et al. Coastal habitats and their importance for the diversity of benthic communities: A species-and trait-based approach. Estuar. Coast. Shelf Sci. 226, 106272. https://doi.org/10.1016/j.ecss.2019.106272 (2019).Article 

Google Scholar 
Loreau, M. et al. Biodiversity and ecosystem functioning: Current knowledge and future challenges. Science 294, 804–808 (2001).ADS 
CAS 

Google Scholar 
Menezes, R. F. et al. Variation in fish community structure, richness, and diversity in 56 Danish lakes with contrasting depth, size, and trophic state: Does the method matter?. Hydrobiologia 710, 47–59 (2013).
Google Scholar 
Hu, M., Wang, C., Liu, Y., Zhang, X. & Jian, S. Fish species composition, distribution and community structure in the lower reaches of Ganjiang River, Jiangxi, China. Sci. Rep. 9, 10100. https://doi.org/10.1038/s41598-019-46600-2 (2019).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Clarke, K. R. & Warwick, R. M. Changes in Marine Communities: An Approach to Statistical Analysis and Interpretation, 2nd ed. (PRIMER-E Ltd, Plymouth Marine Laboratory, 2001).Warwick, R. M. & Clarke, K. R. New biodiversity measures reveal a decrease in taxonomic distinctness with increasing stress. Mar. Ecol. Prog. Ser. 129, 301–305 (1995).ADS 

Google Scholar 
Clarke, K. R. & Warwick, R. M. The taxonomic distinctness measure of biodiversity: Weighting of step lengths between hierarchical levels. Mar. Ecol. Prog. Ser. 184, 21–29 (1999).ADS 

Google Scholar 
Nieto-Navarro, J. T., Zetina-Rejón, M. A., Arreguín-Sánchez, F., Palacios-Salgado, D. & Jordán, F. Changes in fish bycatch during the shrimp fishing season along the eastern coast of the mouth of the Gulf of California. J. Appl. Ichthyol. 29, 610–616 (2013).
Google Scholar 
Escobar-Toledo, F., Zetina-Rejón, M. J. & Duarte, L. O. Measuring the spatial and seasonal variability of community structure and diversity of fish by-catch from tropical shrimp trawling in the Colombian Caribbean Sea. Mar. Biol. Res. 11, 528–539 (2015).
Google Scholar 
Herrera-Valdivia, E., López-Martínez, J., Castillo Vargasmachuca, S. & García-Juárez, A. R. Diversidad taxonómica y funcional en la comunidad de peces de la pesca de arrastre de camarón en el norte del Golfo de California, México. Rev. Biol. Trop. 64, 587–602 (2016).PubMed 

Google Scholar 
Heylings, P., Bensted-Smith, R. & Altamirano, M. Zonificación e historia de la Reserva Marina de Galápagos. In Reserva Marina de Galápagos. Línea Base de la Biodiversidad (eds. Danulat, E. & Edgar, G. J.) 10–21 (Fundación Charles Darwin y Servicio Parque Nacional de Galápagos, 2002).Edgar, G. J. et al. Bias in evaluating the effects of marine protected areas: The importance of baseline data for the Galapagos Marine Reserve. Environ. Conserv. 3, 212–218. https://doi.org/10.1017/S0376892904001584 (2004).Article 

Google Scholar 
Jennings, S., Brierley, A. S. & Walker, J. W. The inshore fish assemblages of the Galápagos archipelago. Biol. Conserv. 70, 49–57 (1994).
Google Scholar 
Brito, A., Pérez-Ruzafaga, A. & Bacallado, J. J. Ictiofauna costera de las islas Galápagos: composición y estructura del poblamiento de los fondos rocosos. Res. Cient. Proy. Galápagos TFCM 5, 61 (1997).
Google Scholar 
Bruneel, S. et al. Assessing the drivers behind the structure and diversity of fish assemblages associated with rocky shores in the Galapagos Archipelago. J. Mar. Sci. Eng. 9, 375. https://doi.org/10.3390/jmse9040375 (2021).Article 

Google Scholar 
Wellington, G. M., Strong, A. E. & Merlen, G. Sea surface temperature variation in the Galápagos Archipelago: A comparison between AVHRR nighttime satellite data and in-situ instrumentation (1982–1998). Bull. Mar. Res. 69, 27–42 (2001).
Google Scholar 
Snell, H., Stone, P. & Snell, H. L. A summary of geographical characteristics of the Galapagos Islands. J. Biogeogr. 23, 619–624 (1996).
Google Scholar 
Bustamante, R. H., et al. Outstanding marine features of Galápagos. In A Biodiversity Vision for the Galapagos Islands: An Exercise for Ecoregional Planning (eds. Bensted-Smith, R. & Dinnerstein, E.) 60–71 (WWF, 2002).Airoldi, L. & Beck, M. W. Loss, status and trends for coastal marine habitats of Europe. In Oceanography and Marine Biology: An Annual Review (eds. Gibson, R. N., Atkinson, R. J. A. & Gordon, J. D. M.) vol. 45, 345–405 (Taylor & Francis, 2007).Carr, M. H., Malone, D. P., Hixon, M. A., Holbrook, S. J. & Schmitt, R. J. How Scuba changed our understanding of nature: underwater breakthrough in reef fish ecology. In Research and Discoveries: The Revolution of Science Through Scuba vol. 39, 157–167 (Smithsonian Contributions to the Marine Sciences, 2013).Durkacz, S. Assessing the Oceanographic Conditions and Distribution of Reef Fish Assemblages Throughout the Galápagos Islands Using Underwater Visual Survey Methods. MSc Thesis (Texas A & M University, 2014).Fischer, W. et al. Guía FAO para la identificación de especies para los fines de pesca. Pacífico Centro-Oriental vol. II–III, 648–1652 (FAO, 1995).Clarke, K. R. & Warwick, R. M. A taxonomic distinctness index and its statistical properties. J. Appl. Ecol. 35, 523–531 (1998).
Google Scholar 
Clarke, K. R. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18, 117–143 (1993).
Google Scholar 
Rosenberg, A., Binford, T. E., Leathery, S., Hill, R. L. & Bickers, K. Ecosystem approaches to fishery management through essential fish habitat. Bull. Mar. Sci. 66, 535–542 (2000).
Google Scholar 
Aburto-Oropeza, O. & Balart, E. F. Community structure of reef fish in several habitats of a rocky reef in the Gulf of California. Mar. Ecol. 22, 283–305 (2001).ADS 

Google Scholar 
Fulton, C. J., Bellwood, D. R. & Wainwright, P. C. Wave energy and swimming performance shape coral reef fish assemblages. Proc. R. Soc. B 272, 827–832 (2005).CAS 
PubMed 

Google Scholar 
Dominici-Arosemena, A. & Wolff, M. Reef fish community structure in the Tropical Eastern Pacific (Panamá): Living on a relatively stable rocky reef environment. Helgol. Mar. Res. 60, 287–305 (2006).ADS 

Google Scholar 
Villegas-Sánchez, C. A., Abitia-Cárdenas, L. A., Gutiérrez-Sánchez, F. J. & Galván-Magaña, F. Rocky-reef fish assemblages at San José Island, Mexico. Rev. Mex. Biodivers. 80, 169–179 (2009).
Google Scholar 
Wiens, J. J. & Graham, C. H. Niche conservatism: Integrating evolution, ecology, and conservation biology. Annu. Rev. Ecol. Evol. Syst. 36, 519–539 (2005).
Google Scholar 
Glynn, P. Some physical and biological determinants of coral community structure in the eastern Pacific. Ecol. Monogr. 46, 431–456 (1976).
Google Scholar 
Ramos-Miranda, J. et al. Changes in four complementary facets of fish diversity in a tropical coastal lagoon after 18 years: A functional interpretation. Mar. Ecol. Prog. Ser. 304, 1–13 (2005).ADS 

Google Scholar 
Gristina, M., Bahri, T., Fiorentino, F. & Garofalo, G. Comparison of demersal fish assemblages in three areas of the Strait of Sicily under different trawling pressure. Fish. Res. 81, 60–71 (2006).
Google Scholar 
Pérez-Ruzafa, A. P., Marcos, C. & Bacallado, J. J. Biodiversidad marina en archipiélagos e islas: patrones de riqueza específica y afinidades faunísticas. Vieraea Folia Scientarum Biologicarum Canariensium. 33, 455–476 (2005).
Google Scholar 
Malcolm, H. A., Jordan, A. & Smith, S. D. Biogeographical and cross-shelf patterns of reef fish assemblages in a transition zone. Mar. Biodivers. 40(3), 181–193 (2010).
Google Scholar 
García-Charton, J. A. & Pérez-Ruzafa, A. P. Correlation between habitat structure and a rocky reef fish assemblage in the Southwest Mediterranean. Mar. Ecol. 19(2), 111–128 (1998).ADS 

Google Scholar 
Mumby, P. J. et al. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427, 533–536 (2004).ADS 
CAS 
PubMed 

Google Scholar 
Unsworth, R. K. F. et al. High connectivity of Indo-Pacific seagrass fish assemblages with mangrove and coral reef habitats. Mar. Ecol. Prog. Ser. 353, 213–224 (2008).ADS 

Google Scholar 
Birkeland, C. & Amesbury, S. S. Fish-transect surveys to determine the influence of neighboring habitats on fish community structure in the tropical Pacific. Co-operation for environmental protection in the Pacific. UNEP Reg. Seas Rep. Stud. 97, 195–202 (1988).
Google Scholar 
Thollot, P., Kulbicki, M., & Wantiez, L. Temporal patterns of species composition in three habitats of the St Vincent Bay area (New Caledonia): Coral reefs, soft bottoms and mangroves. In Proceedings International Soc. Reef Studies. 127–137 (1991).Kulbicki, M. Present knowledge of the structure of coral reef fish assemblages in the Pacific. UNEP Reg. Seas Rep. Stud. 147, 31–53 (1992).
Google Scholar 
Cruz-Romero, M., Chávez, E.A., Espino, E. & García, A. Assessment of a snapper complex (Lutjanus spp.) of the eastern tropical Pacific. In Biology, Fisheries and Culture of Tropical Groupers and Snappers (eds. Arreguín-Sánchez, F., Munro, J. L., Balgos, M. C. & Pauly, D.) 324–330 (ICLARM Conf. Proc. 48, 1996).Aguilar-Santana, F. Biología reproductiva de Prionurus laticlavius (Valenciennes, 1846) (Teleostei: Acanthuridae) en la Costa Sudoccidental del Golfo de California, México. PhD Thesis (Instituto Politécnico Nacional, 2020).Hall, S. The Effects of Fishing on Marine Ecosystems and Communities (Blackwell Science Ltd., 1999).Mangi, S. C. & Roberts, C. M. Quantifying the environmental impacts of artisanal fishing gear on Kenya’s coral reef ecosystems. Mar. Pollut. Bull. 52, 1646–1660 (2006).CAS 
PubMed 

Google Scholar 
Rees, M. J., Jordan, A., Price, O. F., Coleman, M. A. & Davis, A. R. Abiotic surrogates for temperate rocky reef biodiversity: Implications for marine protected areas. Divers. Distrib. 20(3), 284–296 (2014).
Google Scholar 
Ferrari, R. et al. Habitat structural complexity metrics improve predictions of fish abundance and distribution. Ecography 41(7), 1077–1091 (2018).
Google Scholar 
Pihl, L. & Wennhage, H. Structure and diversity of fish assemblages on rocky and soft bottom shores on the Swedish west coast. J. Fish Biol. 61, 148–166 (2002).
Google Scholar 
La Mesa, G., Molinari, A., Gambaccini, S. & Tunesi, L. Spatial pattern of coastal fish assemblages in different habitats in North-western Mediterranean. Mar. Ecol. 32, 104–114 (2011).ADS 

Google Scholar 
Kristensen, L. D. et al. Establishment of blue mussel beds to enhance fish habitats. Appl. Ecol. Environ. Res. 13, 783–798 (2015).
Google Scholar 
Bergström, L., Karlsson, M., Bergström, U., Pihl, L. & Kraufvelin, P. Distribution of mesopredatory fish determined by habitat variables in a predator-depleted coastal system. Mar. Biol. 163, 201. https://doi.org/10.1007/s00227-016-2977-9 (2016).Article 
PubMed 
PubMed Central 

Google Scholar 
Galván-Villa, C. M., Arreola-Robles, J. L., Ríos-Jara, E. & Rodríguez-Zaragoza, F. A. Ensamblajes de peces arrecifales y su relación con el hábitat bentónico de la Isla Isabel, Nayarit, México. Rev. Biol. Mar. Oceanogr. 45, 311–324 (2010).
Google Scholar 
Lunt, J. & Smee, D. L. Turbidity alters estuarine biodiversity and species composition. ICES J. Mar. Sci. 77, 379–387 (2019).
Google Scholar 
Anthony, K. R., Ridd, P. V., Orpin, A. R., Larcombe, P. & Lough, J. Temporal variation of light availability in coastal benthic habitats: Effects of clouds, turbidity, and tides. Limnol. Oceanogr. 49, 2201–2211 (2004).ADS 

Google Scholar 
Helfman, G. S. Patterns of community structure in fishes: Summary and overview’. Environ. Biol. Fishes 3, 129–148 (1978).
Google Scholar 
Helfman, G. S. Fish behaviour by day, night and twilight. In The Behaviour of Teleost Fishes (ed. Pitcher T.J.) (Springer, 1986).Warwick, R. M. & Clarke, K. R. Taxonomic distinctness and environmental assessment. J. Appl. Ecol. 35, 532–543 (1998).
Google Scholar 
Rogers, S. I., Clarke, K. R. & Reynolds, J. D. The taxonomic distinctness of coastal bottom-dwelling fish communities of the North-east Atlantic. J. Anim. Ecol. 68, 769–782 (1999).
Google Scholar 
Robertson, A. I., & Blaber, S. J. M. Plankton, epibenthos and fish communities. In Tropical Mangrove Ecosystems (eds. Robertson, A. I. & Alongi, D. M.) Coastal and Estuarine Studies No. 41, 173–224 (American Geophysical Union, 1992).Koranteng, K. A. Diversity and stability of demersal species assemblages in the Gulf of Guinea. West Afr. J. Appl. Ecol. 2, 49–63 (2001).
Google Scholar 
McCormick, M. I. Comparison of field methods for measuring surface topography and their associations with a tropical reef fish assemblage. Mar. Ecol. Prog. Ser. 112, 87–96 (1994).ADS 

Google Scholar 
Moraes, L. E., Paes, E., Garcia, A., Möller, O. Jr. & Vieira, J. Delayed response of fish abundance to environmental changes: A novel multivariate time-lag approach. Mar. Ecol. Prog. Ser. 456, 159–168 (2012).ADS 

Google Scholar 
Edgar, G. J. et al. El Niño, grazers and fisheries interact to greatly elevate extinction risk for Galapagos marine species. Glob. Change. Biol. 16, 2876–2890 (2010).ADS 

Google Scholar 
Glynn, P. W., Enochs, I. C., Afflerbach, J. A., Brandtneris, V. W. & Serafy, J. E. Eastern Pacific reef fish responses to coral recovery following El Niño disturbances. Mar. Ecol. Prog. Ser. 495, 233–247 (2014).ADS 

Google Scholar  More

Field experiments: collective structuresWe found that self-assembled Eciton hamatum bridges adaptively adjust in response to shifts in the terrain on which they are built. Detailed methods are included in Methods: Field experiments. Briefly, we moved foraging trails onto an apparatus where we could introduce a terrain gap. We repeatedly changed the size of this gap by first incrementally increasing it to 30 mm, by 1 mm every 30 s, and then incrementally contracting it at the same rate (See Fig. 1, Methods: Field experiments, and Supplementary Movie 1). As the size of the gap was expanded (the period before the dotted line in Fig. 2a, b) both the volume and number of ants increased to mean maximum values of 1080 mm3 (standard error, s.e. 84) and 18.9 ants (s.e. 1.6), respectively. Ants typically began forming a bridge when the gap was ~5 mm. As the gap size was decreased (period after dotted line), volume and the number of ants decreased back to zero as ants left the bridge. These broad dynamics across the ten complete trials were similar (Fig. 2a, b, inset panels and Supplementary Figs. 2, 3). Additionally, bridge volume (Fig. 2a) strongly correlated with the number of ants in the bridge (Fig. 2b), indicating that the density of ants per unit volume in these structures is relatively consistent (Pearson correlation coefficients range from 0.88 to 0.98 across the ten trials, see also Supplementary Fig. 4). Bridges broke and quickly reformed in eight of the ten trials; breaks occurred in both experimental phases, and these broken periods were excluded from analyses. Overall, these results show that bridges adjust dynamically to changing terrain geometry, as stretching the bridges caused them to become larger, with more ants, and contracting bridges caused them to become smaller, with fewer ants.Fig. 1: Experimental procedure and data extraction summary.Experiments were conducted on robust E. hamatum foraging trails, which were moved onto the experimental apparatus while it was closed. a Experimental procedure: The size of the gap was increased by 1 mm every 30 s until the gap reached 30 mm (expansion phase), then decreased at the same rate till no gap remained (the contraction phase). b Field setup: Experiments were recorded from both the side and the top, examples of bridges during each phase of the same trial are shown. c Data extraction: Example images and silhouettes from the maximum size bridge (30 mm) of the same trial as the images of 20 mm bridges shown in panel a. The envelopes of the bridges were extracted at a temporal resolution of 1 s; for each focal second, image frames were averaged over 10 s to remove ants walking on the bridge from the extracted envelopes. Envelopes were automatically extracted using hue-saturation-value (HSV) thresholding, with thresholds checked independently for each trial due to lighting differences. Locations of fixed points on the platform were used to re-scale and combine data from the side and top views into a single coordinate system in which 100 pixels = 1 cm. Estimates of bridge volume, mean cross-sectional area, and relative height of the center of mass were recorded from the extracted envelopes as shown. See Methods: Data extraction and Supplementary Note 1 for additional details of the data extraction process, including additional bridge metrics.Full size imageFig. 2: Changes in collective structures in experiments.a, b Volume and group size of self-assembled bridges: a Estimated volume of collective bridge structures over time for one focal trial (main figure) and three other examples (inset). The dotted vertical line indicates the time when the experiment shifted from the expansion phase (increasing gap size) to the contraction phase (decreasing gap size). Gray shading indicates that the bridge was broken or recovering from a break; result metrics may be inaccurate during these periods and they were, therefore, excluded from analyses. b The number of ants in the bridge structure over time for the same focal trial (main figure) and three other examples (inset). c–f Hysteresis: Trials consistently show hysteresis, with bridge status at a particular gap size differing during the expansion and contraction phases, for volume (c), number of ants (d), mean cross-sectional area (e), and tautness, or the height of the center of mass of the bridge from the side view (f; lower values indicate bridge is hanging lower). c–f Panels show result metrics over gap size for the same focal trial as in panels a and b, as well as for three other examples (inset). Points show individual measurements, taken every second, lines are smoothed LOESS (local regression) for the expansion (orange points, dashed orange line) and contraction (green points, solid green line) phases. The area between the smoothed lines (shaded gray) shows the extent of hysteresis. c, e, f) Points are jittered to improve clarity. a–f See Supplementary Figs. 2, 3, 5–8 for all complete trials.Full size imageHowever, these changes were not symmetric—adjustments in the contraction phase were not the inverse of adjustments in the expansion phase. We found consistent hysteresis in several metrics; for a given gap size, bridges were larger and made up of more individuals during the contraction of the gap than the expansion (Fig. 2c, d; t-test for volume: mean extent of hysteresis = 0.43, 95% CI = 0.29 to 0.58, t = 6.7, df = 9, p  More

Model descriptionThe model involves a discrete time, discrete valued stochastic Markov process. Model variables and parameters are given in Tables 1 and 2 respectively. Both time and the number of individuals of each type are constrained to be integer valued. Death and reproductive mutation are stochastic processes derived from sampling from binomial distributions given by the relevant probabilities. All ensemble results give the 100-replicate average for the parameter choices in question.Growth of individuals from species ({S}_{1}) and ({S}_{2}) is proportional to the bio-available level of environmental substances ({R}_{1}) and ({R}_{2}) respectively. At time (t) (where time is in units of biological generations) the change in the number ({N}_{q,j}) of individuals of genotype (j) (non-producer, producer, plastic) within species (q) (({S}_{1}) or ({S}_{2})) can be written as a function of the state of the variables at the previous time-step:$${N}_{q,j}left(t+1right)=left(left({N}_{q,j}left(tright)-{rho }_{q,j}left(tright)right)cdot {G}_{q,j}left(tright)-{{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright)-{delta }_{q,j}left(tright)+{{{{{{rm{{Upsilon }}}}}}}}_{q,xne j}left(tright)right)cdot left(1-frac{{S}_{q}left(tright)}{K}right)$$
(1)
The leftmost bracket on the right-hand side represents the number of individuals escaping starvation (death due to insufficient environmental substance) at the previous time-step and ({G}_{q,j}left(tright)) is the per capita reproductive growth rate. ({{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright)) gives the number of mutant offspring individuals produced during reproduction from parent individuals of genotype (j). ({{{{{{rm{{Upsilon }}}}}}}}_{q,xne j}left(tright)) represents the number of (j) genotype individuals derived from mutation in parent individuals of other genotypes. ({delta }_{q,j}left(tright)) is the number of individuals of genotype (j) lost to random death, and the rightmost bracket relates the total number ({S}_{q}left(tright)) of individuals of species (q) to carrying capacity (K), which represents limitation of growth by any factor other than the relevant environmental substance, e.g. space. (The steady state population size in all simulations shown is below (K) and limited by the environmental substance influx. The carrying capacity is included in the model for computational reasons and as a “crash preventer” but has no qualitative effect on the results).The total number of individuals ({S}_{q}left(tright)) in species (q) is the sum of the number of individuals of each genotype (producer, non-producer and plastic, as discussed in the main text):$${S}_{q}left(tright)=mathop{sum }limits_{j=1}^{{j}_{{total}}}{N}_{q,j}left(tright)={N}_{q,{prod}}left(tright)+{N}_{q,{non}-{prod}}left(tright)+{N}_{q,{plast}}left(tright)$$
(2)
The genotype-specific reproductive growth rate ({G}_{q,j}left(tright)) (again for genotype (j) within species (q), time (t)), gives the number of offspring individuals produced per parent individual, per time-step. Growth rate is an increasing function of the bio-available level of environmental substance ({R}_{q,{BIOAVAIL}{ABLE}}) (the subscript (q) being identical because species ({S}_{1}) and ({S}_{2}) assimilate substances ({R}_{1}) and ({R}_{2}) respectively). Growth rate also includes a substance-to-biomass conversion efficiency parameter ({f}_{{conv}}) and a genotype-specific per capita term ({G}_{q,j,{PR}}) (number of offspring per parent, per unit environmental substance assimilated, per unit time). In the absence of growth-limitation by environmental substance levels, growth rate is capped at a genotype-specific maximum ({G}_{q,{jMAX}}):$${G}_{q,j}left(tright)={MIN}[{G}_{q,j,{PR}}cdot {R}_{q,{BIOAVAILABLE}}(t)cdot {f}_{{conv}},{G}_{q,{jMAX}}]$$
(3)
$${G}_{q,{non}-{prod},{PR}}={G}_{0}$$
(4)
$${G}_{q,{non}-{prod},{MAX}}={G}_{0}cdot {R}_{{assimMAX}}$$
(5)
({G}_{0}) is the baseline number of offspring, per parent, per unit substance assimilated. ({R}_{{assimMAX}}) is a universal maximum potential number of units of environmental substance that can be assimilated by a single individual per time-step (i.e. representing basic physiological constraints on growth). The producer genotype incurs a per capita reproductive growth rate cost ({kappa }_{{prod}}) relative to the non-producer:$${G}_{q,{prod},{PR}}={G}_{0}cdot (1-{kappa }_{{prod}})$$
(6)
$${G}_{q,{prod},{MAX}}={G}_{0}cdot (1-{kappa }_{{prod}})cdot {R}_{{assimMAX}}$$
(7)
This growth rate formulation is therefore a highly simplified linearization of the Michaelis-Menten kinetics normally used in models of resource and nutrient assimilation.The plastic genotype switches phenotype depending upon the level of environmental substance relative to a fixed threshold ({{R}_{q,{BIOAVAILABLE}}}_{{crit}}), in effect becoming a second non-producer genotype below this threshold and a second producer genotype above it:$${IF}[{R}_{q,{BIOAVAILABLE}}(t)ge {{R}_{q,{BIOAVAILABLE}}}_{{crit}}],{G}_{q,{plast}}left(tright)={G}_{q,{prod}}left(tright)$$$${ELSEIF}[{R}_{q,{BIOAVAILA}{BLE}}left(tright) , < , {{R}_{q,{BIOAVAILABLE}}}_{{crit}}],{G}_{q,{plast}}left(tright)={G}_{q,{non}-{prod}}left(tright)$$ (8) There is no spatial structure whatsoever, thus access to environmental substance is uniform across individuals. The bioavailable quantity of each environmental substance is simply the total amount ({R}_{q,{NET}}(t)) divided by the total number of individuals assimilating it:$${R}_{q,{BIOAVAILABLE}}left(tright)=frac{{R}_{q,{NET}}(t)}{{S}_{q}left(tright)}$$ (9) We allow the per capita reproductive growth rate to fall below ({G}_{q,j}left(tright)=1), which, if interpreted deterministically at the individual level would correspond to an individual failing to sustain its biomass to the next time-step and thus dying. However, a population-level average ({G}_{q,j}left(tright) , < , 1) is interpretable in terms of a thinning factor that maps between discretized individuals and continuously distributed environmental substance. Thus, a thinning factor of (left(1-{G}_{q,j}left(tright)right)) is used to calculate the total number of individuals dying of starvation ({rho }_{q,j}) (again genotype (j), species (q)). This represents pre-reproduction deaths, corresponding to the difference between the actual population size and the population size that the environmental substance pool is capable of supporting. ({rho }_{q,j}left(tright)) is constrained to be an integer and is zero for ({G}_{q,j}left(tright) , > , 1):$${rho }_{q,j}left(tright)={N}_{q,j}left(tright)cdot {MAX}left[0,left(1-{G}_{q,j}left(tright)right)right]$$
(10)
A subset of offspring are a different genotype from their parent via mutation. For parent genotype (j), the number ({{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright)) of mutant offspring with genotype (ne j) is calculated using baseline mutation probability per reproductive event ({mu }_{0}), with the total number of new individuals produced by the parent individuals surviving starvation ({G}_{q,j}left(tright)cdot left({N}_{q,j}left(tright)-{rho }_{q,j}left(tright)right)). The total number of mutant offspring ({{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright)) is thus a binomially distributed random variable with success probability ({mu }_{0}) and number of trials ({G}_{q,j}left(tright)cdot left({N}_{q,j}left(tright)-{rho }_{q,j}left(tright)right)). The expected value (Eleft[{{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright)right]) is the product of these two numbers:$$ {{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright) sim Bleft({G}_{q,j}left(tright)cdot left({N}_{q,j}left(tright)-{rho }_{q,j}left(tright)right),{mu }_{0}right), \ Eleft[{{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright)right]={G}_{q,j}left(tright)cdot left({N}_{q,j}left(tright)-{rho }_{q,j}left(tright)right)cdot {mu }_{0}$$
(11)
Mutation to genotype (j) from the other genotypes is calculated in exactly the same way using the number and reproductive growth rates of the relevant (other) genotypes. Any particular mutant offspring is randomly allocated to one of the other genotypes with equal probability ({p}_{kto j}=frac{1}{{j}_{{total}}-1}=0.5) (where ({j}_{{total}}=3) is the total number of genotypes per species). The expected number of offspring with genotype (j) produced by mutation within parent offspring of other genotypes (kne j) is therefore:$$E[{{{{{{rm{{Upsilon }}}}}}}}_{q,x , ne , j}left(tright)]={left(mathop{sum }limits_{k=1}^{{k}_{{to}{tal}}}{{{{{{rm{{Upsilon }}}}}}}}_{q,k}left(tright)right)}_{kne j}cdot frac{1}{{j}_{{total}}-1}$$
(12)
Independently of reproduction and assimilation of environmental substance, any given individual has a probability ({delta }_{0}) at each time point of death due to stochastic factors. The genotype/species specific number of such deaths is again a random sample from a binomial distribution, with success probability ({delta }_{0}):$${delta }_{q,j}left(tright) sim Bleft({N}_{q,j}left(tright),{delta }_{0}right),Eleft[{delta }_{q,j}left(tright)right]={N}_{q,j}left(tright)cdot {delta }_{0}$$
(13)
The net quantity of growth-limiting environmental substance at each time-step is given by the difference between total biotic assimilation ({A}_{{R}_{q}}) and the production ({P}_{{R}_{q}}) and abiotic input ({varphi }_{{R}_{q}}) fluxes:$${R}_{q,{NET}}(t+1)={varphi }_{{R}_{q}}(t)+{P}_{{R}_{q}}(t)-{A}_{{R}_{q}}(t)$$
(14)
The abiotic net influx is the sum of two fluxes. First, an input term that is the product of a baseline scaling factor ({{varphi }_{0}}_{{R}_{q}}) and a model forcing (frac{partial {t}_{{geo}}}{partial {t}_{{bio}}}) representing the mapping between abiotic-geological and biotic-evolutionary timescales. In practice (frac{partial {t}_{{geo}}}{partial {t}_{{bio}}}(t)) was set to either (1) or (0) or (in fluctuation runs) a time-dependent switching between the two. (More sophisticated implementations of (frac{partial {t}_{{geo}}}{partial {t}_{{bio}}}(t)), e.g. sinusoidal oscillations and stochastic time dependence, were attempted but made little qualitative difference to the results). Second, an abiotic removal term that scales linearly with the quantity of environmental substance:$${varphi }_{{R}_{q}}(t)={{varphi }_{0}}_{{R}_{q}}cdot frac{partial {t}_{{geo}}}{partial {t}_{{bio}}}left(tright)-frac{{R}_{q,{NET}(t)}}{{R}_{q,{NET}0}}$$
(15)
where ({R}_{q,{NET}0}) is a normalization factor representing the sensitivity of the abiotic efflux to the influx. In the absence of any biota and for (frac{partial {t}_{{geo}}}{partial {t}_{{bio}}}=1) the steady state environmental substance level is immediately given by (15) as ({R}_{q,{NET}(t)}={R}_{q,{NET}0}cdot {{varphi }_{0}}_{{R}_{q}}), thus the numerical value of ({R}_{q,{NET}0}) corresponds to the abiotic steady state residence time.Total biotic assimilation ({A}_{R}) of each environmental substance is given by:$${A}_{{R}_{q}}left(tright)=mathop{sum }limits_{k=1}^{{j}_{{total}}}frac{{G}_{q,k}left(tright)cdot left({N}_{q,k}left(tright)-{rho }_{q,k}left(tright)right)}{{G}_{q,{kPR}}}$$
(16)
The numerator gives the total number of individuals produced as a result of biological assimilation of environmental substance ({R}_{q}) and the denominator is the genotype specific number of individuals produced per unit substance assimilated, dividing through by which therefore converts to total units of substance assimilated by the population as a whole.Net biotic ({P}_{{R}_{q}{NET}}) production of substance ({R}_{q}) by the producer genotype in the other species (p) is calculated equivalently, via the product of the per capita production rate ({P}_{q,{prod}}) and the total number of reproducing individuals:$${P}_{q,{prod}}left(tright)=frac{{MIN}[{G}_{q,{prod}}left(tright)cdot {f}_{{convprod}},{G}_{q,{prodMAX}}]}{{G}_{p,{PRODUCER;PR}}}$$
(17)
$${P}_{{R}_{q}{NET}}left(tright)={P}_{q,{prod}}left(tright)cdot left({N}_{p,{PRODUCER}}left(tright)-{rho }_{p,P{RODUCER}}left(tright)right)$$
(18)
Where ({f}_{{conv},{PROD}}) is the per capita efficiency by which producers convert the environmental substance that they assimilate into the by-product they produce (note that the equivalent conversion efficiency for assimilation ({f}_{{conv}}) already appears in the growth functions of each genotype, therefore does not appear in Eq. (17)).The residence time ({T}_{{R}_{q}})of each environmental substance is given by the net quantity of this substance divided by the influx, to give units of the average number of biological generations a unit of environmental substance spends in the relevant pool before being removed. In those simulations in which the abiotic influx ({varphi }_{{R}_{q}}(t)) was set to zero (i.e. during the shut-off intervals) production ({P}_{{R}_{q}{NE}T}left(tright)) was used as an alternative denominator:$${IF}left[{varphi }_{{R}_{q}}left(tright) , > , 0,{T}_{{R}_{q}}left(tright)=frac{{R}_{q,{NET}}left(tright)}{{varphi }_{{R}_{q}}left(tright)}right]!,{IF}left[left(left({varphi }_{{R}_{q}}left(tright)=0right){& }left({P}_{{R}_{q}{NET}}left(tright) , > ,0right)right)!,{T}_{{R}_{q}}left(tright)=frac{{R}_{q,{NET}}left(tright)}{{P}_{{R}_{q}{NET}}left(tright)}right]{ELSE}[{T}_{{R}_{q}}left(tright)=0]$$
(19)
The cycling ratio ({{CR}}_{{R}_{q}}) of each substance is given by the ratio between net biotic assimilation of that substance ({A}_{{R}_{q}}left(tright)) and the abiotic influx of that substance ({varphi }_{{R}_{q}}left(tright)). As with the residence time, when the abiotic influx was zero, the input from biological production was used as an alternative denominator:$${IF}left[{varphi }_{{R}_{q}}left(tright) , > , 0,{{CR}}_{{R}_{q}}left(tright)=frac{{A}_{{R}_{q}}left(tright)}{{varphi }_{{R}_{q}}left(tright)}right]{IF}left[left(left({varphi }_{{R}_{q}}left(tright)=0right){{& }}left({P}_{{R}_{q}{NET}}left(tright) , > ,0right)right),{{CR}}_{{R}_{q}}left(tright)=frac{{A}_{{R}_{q}}left(tright)}{{P}_{{R}_{q}{NET}}left(tright)}right]{ELSE}[{{CR}}_{{R}_{q}}left(tright)=0]$$
(20)
Deterministic approximation to steady stateAssume that at steady state substance assimilation will reach a maximal state such that the level of environmental substance is limiting to population size. Assume that such a state is below the level ({{R}_{q,{BIOAVAILABLE}}}_{{crit}}) at which the plastic genotype effectively becomes a second non-producer genotype and can thus be subsumed into non-producer frequency, such that (2) becomes ({S}_{q}left(tright)=mathop{sum }nolimits_{j=1}^{{j}_{{total}}}{N}_{q,j}left(tright)={N}_{q,{prod}}left(tright)+{N}_{q,{non}-{prod}}left(tright)). Assume that there are non-zero starvations at each time-step for all genotypes, which implies growth rate ({G}_{q,j}left(tright) , < , 1,ll {G}_{q,{jMAX}},forall j,q), which gives by (3)({G}_{q,j}left(tright)={G}_{q,j,{PR}}cdot {R}_{q,{BIOAVAILABLE}}left(tright)cdot {f}_{{conv}}). Substituting this into (10), then the first bracketed term in (1), then labeling the post-starvation number of individuals as ({({N}_{q,j}left(tright))}_{{NET}}):$${({N}_{q,j}left(tright))}_{{NET}}=left({N}_{q,j}left(tright)-{rho }_{q,j}left(tright)right)={N}_{q,j}left(tright)cdot left(1-left(1-{G}_{q,j}left(tright)right)right)={N}_{q,j}left(tright)cdot {G}_{q,j}left(tright)$$ (21) Approximate (14) deterministically by a fixed fractional parameter corresponding to the baseline random death rate:$${delta }_{q,j}left(tright)approx {N}_{q,j}left(tright)cdot {delta }_{0}$$ (22) Doing the same for mutation:$${{{{{{rm{{Upsilon }}}}}}}}_{q,j}left(tright)={G}_{q,j}left(tright)cdot {left({N}_{q,j}left(tright)right)}_{{NET}}cdot {mu }_{0}={N}_{q,j}left(tright)cdot {{G}_{q,j}left(tright)}^{2}cdot {mu }_{0}$$ (23) The term in (12) for mutation to (j) from other genotypes simplifies to$${{{{{{rm{{Upsilon }}}}}}}}_{q,x ,ne , j}left(tright)={sum }_{k,=,1}^{{j}_{{total}}}frac{{G}_{q,k}left(tright)cdot {left({N}_{q,k}left(tright)right)}_{{NET}}cdot {mu }_{0}}{{j}_{{total}}-1}={N}_{q,k}left(tright)cdot {{G}_{q,k}left(tright)}^{2}cdot {mu }_{0}$$Substituting Eqs. (21–23) into (1):$${N}_{q,j}left(tright)cdot {{G}_{q,j}left(tright)}^{2}cdot (1-{mu }_{0})-{N}_{q,j}left(tright)cdot {delta }_{0}+{N}_{q,k}left(tright)cdot {{G}_{q,k}left(tright)}^{2}cdot {mu }_{0}=0$$ (24) Noting that by Eqs. (3)–(5) combined with the above assumptions, the growth rate of the producer can be written as:$${G}_{q,{prod}}left(tright)={G}_{q,{cheat}}left(tright)cdot (1-{kappa }_{{prod},q})$$ (25) Writing (24) explicitly for each genotype:$${N}_{q,{non}-{prod}}left(tright)cdot {{G}_{q,{non}-{prod}}left(tright)}^{2}cdot (1-{mu }_{0})-{N}_{q,{non}-{prod}}left(tright)cdot {delta }_{0}+{N}_{q,{prod}}left(tright)cdot {{G}_{q,{prod}}left(tright)}^{2}cdot {mu }_{0}=0$$$${N}_{q,{prod}}left(tright)cdot {{G}_{q,{prod}}left(tright)}^{2}cdot (1-{mu }_{0})-{N}_{q,{prod}}left(tright)cdot {delta }_{0}+{N}_{q,{non}-{prod}}left(tright)cdot {{G}_{q,{non}-{prod}}left(tright)}^{2}cdot {mu }_{0}=0$$Adding:$${N}_{q,{non}-{prod}}left(tright)cdot {{G}_{q,{non}-{prod}}left(tright)}^{2}cdot left(1-{mu }_{0}right)-{N}_{q,{non}-{prod}}left(tright)cdot {delta }_{0}$$$$+{N}_{q,{prod}}left(tright)cdot {left({G}_{q,{non}-{prod}}left(tright)cdot left(1-{kappa }_{{prod},q}right)right)}^{2}cdot {mu }_{0}+{N}_{q,{prod}}left(tright)cdot {left({G}_{q,{non}-{prod}}left(tright)cdot left(1-{kappa }_{{prod},q}right)right)}^{2}cdot left(1-{mu }_{0}right)$$$$-{N}_{q,{prod}}left(tright)cdot {delta }_{0}+{N}_{q,{non}-{prod}}left(tright)cdot {{G}_{q,{non}-{prod}}left(tright)}^{2}cdot {mu }_{0}=0$$Because the mutation terms cancel:$$left({N}_{q,{non}-{prod}}left(tright)+{N}_{q,{prod}}left(tright)cdot {left(1-{kappa }_{{prod},q}right)}^{2}right)cdot left({{G}_{q,{non}-{prod}}left(tright)}^{2}-{delta }_{0}right)=0$$ (26) Substituting in for the growth rate terms (3–7) gives:$$left({N}_{q,{non}-{prod}}(t)+{N}_{q,{prod}}(t)cdot {left(1-{kappa }_{{prod},q}right)}^{2}right)cdot left({left({G}_{0}cdot {R}_{q,{BIOAVAILABLE}}(t)cdot {f}_{{conv}}right)}^{2}-{delta }_{0}right)=0$$ (27) By (16), (4), (6) and the above, total steady state assimilation of the growth limiting environmental substance by species (q) is:$${A}_{{R}_{q}}left(tright)=mathop{sum }limits_{k=1}^{{j}_{{total}}}frac{{G}_{q,k}left(tright)cdot left({N}_{q,k}left(tright)-{rho }_{q,k}left(tright)right)}{{G}_{q,{kPR}}}$$$$kern2.4pc=frac{{N}_{q,{non}-{prod}}left(tright)cdot {left({G}_{q,{non}-{prod}}left(tright)right)}^{2}}{{G}_{0}}+frac{left({N}_{q,{prod}}left(tright)cdot {left(1-{kappa }_{{prod},q}right)}^{2}right)cdot {left({G}_{q,{non}-{prod}}left(tright)right)}^{2}}{{G}_{0}cdot left(1-{kappa }_{{prod}}right)}$$$$=left({N}_{q,{non}-{prod}}left(tright)+{N}_{q,{prod}}left(tright)cdot left(1-{kappa }_{{prod},q}right)right)cdot {{G}_{0}cdot ({R}_{q,{BIOAVAILABLE}}left(tright)cdot {f}_{{conv}})}^{2}$$ (28) By (16–18), production of this substance by the producer allele in the other species (p , ne , q), assuming the various arguments above simultaneously apply to this species, is:$${P}_{{R}_{q}}left(tright)= frac{{G}_{p,{prod}}left(tright)cdot left({N}_{p,{prod}}left(tright)-{rho }_{p,{PRODUCER}}left(tright)right)}{{G}_{p,{prodPR}}}cdot {f}_{{conv},{PROD}}\ = {N}_{p,{prod}}left(tright)cdot left(1-{kappa }_{{prod},p}right) cdot {G}_{0}cdot {({R}_{p,{BIOAVAILABLE}}left(tright)cdot {f}_{{conv}})}^{2}cdot {f}_{{conv},{PROD}}$$ (29) Balance between input and output fluxes of each environmental substance requires ({varphi }_{{R}_{q}}left(tright)+{P}_{{R}_{q}}left(tright)={A}_{{R}_{q}}left(tright)), meaning that by substituting in ({A}_{{R}_{q}}left(tright)) from (28) it is possible to solve for bioavailable substance level, then substitute in the production flux of this substance derived from the producer allele in the other species (p,ne, q):$${R}_{q,{BIO}{AVAILABLE}}left(tright) =frac{1}{{f}!_{{conv}}}sqrt{frac{{varphi }_{{R}_{q}}left(tright)+{P}_{{R}_{q}}left(tright)}{left({N}_{q,{non}-{prod}}left(tright)+{N}_{q,{prod}}left(tright)cdot left(1-{kappa }_{{prod},q}right)right)cdot {G}_{0}}}\ =frac{1}{{f}!_{{conv}}}sqrt{frac{{varphi }_{{R}_{q}}left(tright)+{N}_{p,{prod}}left(tright)cdot left(1-{kappa }_{{prod},p}right)cdot {G}_{0}cdot {({R}_{p,{BIOAVAILABLE}}left(tright)cdot {f}_{{conv}})}^{2}cdot {f}_{{conv},{PROD}}}{left({N}_{q,{non}-{prod}}left(tright)+{N}_{q,{prod}}left(tright)cdot left(1-{kappa }_{{prod},q}right)right)cdot {G}_{0}}}$$ (30) Substituting this into (27) gives a symmetrical condition for steady state genotype frequencies and substance levels across the system:$$ left({N}_{q,{non}-{prod}}left(tright)+{N}_{q,{prod}}left(tright)cdot {left(1-{kappa }_{{prod},q}right)}^{2}right)cdot\ left(frac{{varphi }_{{R}_{q}}left(tright)+{N}_{p,{prod}}left(tright)cdot left(1-{kappa }_{{prod},p}right)cdot {G}_{0}cdot {({R}_{p,{BIOAVAILABLE}}left(tright)cdot {f}_{{conv}})}^{2}cdot {f}_{{conv},{PROD}}}{left({N}_{q,{non}-{prod}}left(tright)+{N}_{q,{prod}}left(tright)cdot left(1-{kappa }_{{prod},q}right)right)cdot {G}_{0}}-{delta }_{0}right)=0$$ (31) This solution illustrates the intuitive ideas that growth and reproduction balance random death at steady state and that the associated producer frequency is lower than that of the non-producer by a factor of the cost. (This factor is of second order because the growth rate is used both directly and (by (10)) in the calculation of starvations). Because our model is a discrete stochastic process, (31) can be viewed as an approximation to a steady state condition, subject to the above assumptions combined with the continuous generation of producers by mutation at a sufficient rate to preclude their extinction. The key point is that over long timescales in the finite populations with which we deal, organism-level selection unavoidably favors the non-producer, with no possibility for multi-level fecundity selection. The producer’s stable presence is thus attributable to the combination of mutation and cycle-level selection. More

Continental, regional, and local spatiotemporal patterns of carbon lossFor Africa’s primary tropical humid forest, carbon losses due to forest disturbances reached 42.2 ± 5.1 MtC yr−1 (mean ± standard deviation, where MtC yr−1 is one million metric tons of carbon loss per year) in 2019 and 53.4 ± 6.5 MtC yr−1 in 2020. Just 9 countries out of the 23 analyzed accounted for 95.0% of total gross losses in 2019 and 94.3% in 2020. These countries contain about 95.7% of all primary tropical humid forests of Africa, with the DRC accounting for 52.8%, Gabon 11.8%, the Republic of the Congo 11.0%, and Cameroon 9.8%. Of these, DRC and Cameroon were responsible for 49.3% and 19.1% of losses in 2019 and 44.7% and 20.6% in 2020. DRC and Cameroon had an annual increase of 15.0% and 36.5% respectively, between 2019 and 2020. From countries with at least 1 MtC emitted in the two years analyzed, Madagascar had the highest annual increase in carbon loss (+153.9%), while Equatorial Guinea is the only country with a decrease in carbon loss (−20.1%). Extending the carbon loss analysis for both past and future will help to better understand these variations and whether the COVID-19 global pandemic had any influence on the general increase between 2019 and 202019. While the absolute numbers for carbon loss estimates should be treated carefully and a sample-based approach should be preferred for an unbiased estimate of absolute numbers20, we focused our analysis on the trends of carbon loss at the continental, country, and local scale (Fig. 1 and Supplementary Fig. 1).Fig. 1: Carbon loss across Africa’s rainforests.We analyzed 23 countries containing primary moist forest. The aboveground carbon stock (green palette) underlies the carbon loss estimations (red palette). Several hotspots can be seen across these regions. The uncertainties of the carbon loss estimations are expressed as standard deviations and shown in Supplementary Fig. 1.Full size imageThe high temporal detail of the analysis revealed various monthly patterns of carbon losses for countries, highly related to local rainfall patterns18 (Fig. 2). Countries like Cameroon, Liberia, Nigeria, Central African Republic (CAR), and Madagascar showed a clear dry-wet seasonal variation in carbon loss per year, while the Republic of the Congo and the DRC, due to their latitudinal extent, exhibited two dry-wet season variations per year with varying intensities (Fig. 2). The seasonal variation can be explained by higher accessibility to forests during the dry months when activities related to smallholder agriculture and logging are more feasible than in the wet season when many roads become inaccessible.Fig. 2: Temporal patterns of carbon loss for the top 10 countries.We show monthly statistics for 2019 and 2020 and the associated uncertainty (black lines). We separate between high (red bars) and low (yellow bars) confidence alerts, the latter showing up for the last 3 months of 2020.Full size imageOne of the highest differences between the months with the most and the least carbon losses was found for Madagascar (72 times more carbon loss in March compared to November 2019). In CAR, the three consecutive months with the highest cumulative carbon loss (January to March 2020) contributed to 75.7% of the total annual loss (between February and April 2020), in Nigeria 73.9% (January to March 2020), Liberia 73.1% (February to April 2020), Madagascar 70.7% (September to November 2020), and Cameroon 62.2% (January to March 2020). Lower percentages were found for countries with mixed seasonality and patterns, like DRC 36.7% (January to March 2020), and the Republic of the Congo 32.8% (January to March 2020) (Fig. 2). For the latter two countries, we expect better-defined peaks of carbon loss at local scales, where climatic conditions are not mixed. The annual cumulative carbon loss (%) per country (Fig. 3) showed that Liberia, Nigeria, CAR, and Cameroon reached between 70-90% of their annual carbon loss in April, while Madagascar reached 60% in October. The DRC, Gabon, Republic of the Congo, Equatorial Guinea, and Ghana have a more gradual monthly increase of cumulative carbon loss with less contrasting seasonality effects. Monthly patterns of carbon losses between the two years analyzed resulted in a correlation coefficient of 0.94 for the CAR, 0.92 for the DRC, 0.91 for Madagascar, 0.90 for Gabon, and 0.83 for Cameroon (Supplementary Fig. 2). For the Republic of the Congo, the two years correlated 0.51. Knowing the peak months of carbon loss for each country and that these patterns are repeatable from one year to another can contribute to better target and prioritize enforcement activities, as well as predicting future patterns and early reporting of annual forest carbon losses.Fig. 3: Annual cumulative carbon loss (%) for both years analyzed, 2019 and 2020.Africa’s total cumulative carbon loss is shown with a black line. The 10 topmost emitting countries out of 23 countries analyzed are shown and represented by distinct colored lines.Full size imageSeveral hotspots of carbon losses can be seen in Fig. 1. The high spatial and temporal details of our analysis are shown in Fig. 4, where several local examples with different drivers of forest disturbances are shown, like logging roads, selective logging, mining, oil palm plantations, urban expansion, and small-holder agriculture. This kind of information, coupled with auxiliary datasets (e.g., legal concessions, protected areas) can identify the legality of forest disturbance21.Fig. 4: Local examples of approx. 10 × 10 km in extent showing different spatiotemporal patterns and drivers of carbon loss.The first column shows the carbon loss, the second column the associated uncertainty, the third column the day-of-the-year when the loss occurred, and the last column shows the monthly distribution of carbon loss and associated uncertainty for each local example. The center coordinates of each location are shown in the third column as latitude and longitude. Exact locations are shown in Supplementary Fig. 3. a Logging roads and selective logging in the Central African Republic, b mining of gold and titanium in the Republic of the Congo, c development of an oil palm plantation in Cameroon, d forest disturbance related to building a new capital city in Equatorial Guinea, and e small-scale agriculture expansion at the edge of the forest in the DRC.Full size imageImplications of rapid monitoring of local carbon lossNear-real time alerts combined with biomass maps result in spatially explicit forest carbon loss, unlike global tabular statistics of national data22,23. We provide new insights into the spatiotemporal dynamics of carbon loss with consistent assessment of accuracy that could enable transparency and completeness for countries reporting on their REDD + progress to the UNFCCC24. We provide monthly carbon loss estimates that could play a key role in local, national, and international forest initiatives for global carbon policy goals25. Such a system can be implemented with minimal costs and is based on open-source datasets and Google Earth Engine cloud computing platform26, thus enabling cost-effective national monitoring of forest carbon loss7. Providing rapid reporting on the location, time, and amount of carbon lost across Africa’s primary humid forest will help undertake immediate action to protect and conserve carbon-rich threatened forests. Furthermore, countries will be able to predict and estimate their annual carbon loss before a reporting period ends, thus having the opportunity to adjust their practices to meet their country-specific commitments for climate change mitigation initiatives.Limitations and future improvementsWe used the RADD alerts (Radar for Detecting Deforestation)18 with a minimum mapping unit (MMU) of 0.2 ha as accuracy estimates were available for this MMU. Events smaller than 0.2 ha would add to the total carbon loss but are by nature associated with higher uncertainties18. The implications of the RADD alerts using a global humid tropical forest product as a forest baseline for 201816,27,28 are twofold. First, the global nature of this product might result in inconsistencies at the local level18. Second, because the forest cover loss information used to generate the forest baseline is based on optical Landsat data, persistent cloud cover in the second half of 2018 in some areas led to missed reporting of forest disturbances, thus being detected at the beginning of 2019 by the RADD alerts. This possible overestimation of carbon loss at the start of 2019 is not an issue for a near-real-time alerting system since later months are not affected. Furthermore, the alerts do not distinguish between human-induced disturbances and natural forest disturbances18. When a new forest disturbance alert is detected, it will be confirmed or rejected within 90 days by subsequent Sentinel-1 images18. That is why our carbon loss reporting separates between high and low confidence alerts for the last three months of 2020, which is common for most forest disturbance alerting products18,29. We separated all the alerts into core and boundary pixels. Core alerts represent complete tree cover removal and we assumed complete carbon loss within a pixel. For boundary alerts, we assumed a 50% carbon loss since these mainly represent forest disturbances with partial tree cover removal. Detecting and quantifying the level of degradation remains challenging and future developments will minimize this uncertainty by providing variable percentages of degraded forest30. The timeliness and spatial details of future forest disturbance alerting products will improve with the availability of open access long-wavelength radar data from near-future satellite missions (e.g., NISAR L-band SAR in 202331), by using a combination of optical and radar forest disturbance alert products, and integration with high-resolution satellite products.We relied on an aboveground biomass baseline map from 201832, prior to RADD alerts starting from 2019. Biomass estimation for the tropical moist forests is based on ALOS-2 PALSAR-2 L-band satellite and its usage needs to account for the local biases, especially underestimating AGB values higher than 250 Mg ha−1 (ref. 32). Although we reduced this underestimation by adjusting the AGB map based on ground field data, more research is needed on providing up-to-date high-resolution aboveground carbon estimates33 that could further increase the accuracy of local carbon loss estimation. Radar-based estimation of forest carbon stocks is challenging over mountainous terrain and is less accurate in complex canopies3 and future integration of radar and optical satellite data will provide more robust estimates33. Nevertheless, new spaceborne missions (e.g., GEDI34, BIOMASS35) will provide an unprecedented amount of forest structure samples that will improve the algorithms and thus the final accuracy of aboveground biomass estimates.We focused on exploring and analyzing local carbon losses and showing high temporal and spatial patterns of carbon losses. We showed the country statistics to emphasize the temporal dynamics of carbon losses and compare the temporal profiles across our study region. Our approach was not to provide stratified area estimations36 associated with forest disturbances but we used this concept in the sense that we had a stratified sample of higher quality reference data18 to estimate the omission and commission errors and consider those in our uncertainty estimation on the pixel level. The analysis showed that omission and commission errors are small and rather balanced, and thus do not result in a major area bias for the forest disturbances. The uncertainties of the aboveground biomass product32 were adjusted for known regional biases using regional forest biomass plot data sources. With this approach, the original aboveground biomass map bias was partly corrected using a model-based approach deemed to be an alternative to a sample-based approach whenever country data are unavailable37. Our uncertainty analysis and error reduction showed that we expect only minor bias in the forest disturbance and the biomass data and the remaining uncertainties are propagated in our pixel-based uncertainty layer. More