Warming, deoxygenation and extinction mechanisms in the oceans

The fossil record is a reliable archive for testing biological hypotheses of faunal change and extinction against the backdrop of rapid, intense climate change (Barnosky et al., Reference Barnosky, Hadly, Bascompte, Berlow, Brown, Fortelius, Getz, Harte, Hastings, Marquet, Martinez, Mooers, Roopnarine, Vermeij, Williams, Gillespie, Kitzes, Marshall, Matzke, Mindell, Revilla and Smith2012; Vermeij and Roopnarine, Reference Vermeij and Roopnarine2013; Orzechowski et al., Reference Orzechowski, Lockwood, Byrnes, Anderson, Finnegan, Finkel, Harnik, Lindberg, Liow, Lotze, Mcclain, Mcguire, O’Dea, Pandolfi, Simpson and Tittensor2015; Fastovich et al., Reference Fastovich, Meyers, Saupe, Williams, Dornelas, Dowding, Finnegan, Huang, Jonkers, Kiessling, Kocsis, Li, Liow, Na, Penny, Pippenger, Renaudie, Rillo, Smith, Steinbauer, Sugawara, Tomašových, Yasuhara and Hull2025). Global warming, in particular, has been a recurrent factor in biodiversity crises. As case studies, we will start by illustrating what happened at the end of the Permian, in coincidence with the largest mass extinction ever (Benton, Reference Benton2018), and during the early, greenhouse Earth period of the Cenozoic, in correspondence with Paleogene hyperthermals (Bowen et al., Reference Bowen, Bralower, Delaney, Dickens, Kelly, Koch, Kump, Meng, Sloan, Thomas, Wing and Zachos2006). Fuller reviews specifically focusing on these events are available elsewhere and cited below. We direct the reader toward these valuable sources to delve deeper into the geochemistry, paleontology and climatology of the events.

The Permian/Triassic mass extinction, colloquially known as the ‘Great Dying’, wiped out around 80–90% of marine animal species (Stanley, Reference Stanley2016; Finnegan et al., Reference Finnegan, Harnik, Lockwood, Lotze, McClenachan and Kahanamoku2024). Initiated by massive volcanic activity recorded in the Siberian Traps, the volcanism released vast amounts of greenhouse gases (CO₂, CH₄) and sulfur dioxide, driving atmospheric carbon dioxide up to ten times as much as today (that is ~4,000 ppm, Cui and Kump, Reference Cui and Kump2015). The primary kill mechanisms were the extreme temperatures, with warming topping >10 °C increase from pre-extinction times (Penn et al., Reference Penn, Deutsch, Payne and Sperling2018; Gliwa et al., Reference Gliwa, Wiedenbeck, Schobben, Ullmann, Kiessling, Ghaderi, Struck and Korn2022), and ocean acidification and deoxygenation (oxygen-starved oceans, Hülse et al., Reference Hülse, Lau, van de Velde, Arndt, Meyer and Ridgwell2021). New research indicates that the crisis was triggered by Mega El Niño events linked to the collapse of the meridional overturning circulation (Sun et al., Reference Sun, Farnsworth, Joachimski, Wignall, Krystyn, Bond, Ravidà and Valdes2024b). On land, the environmental crisis manifested through profound disruption of terrestrial ecosystems, including widespread vegetation loss, increased wildfire activity, enhanced soil erosion and major restructuring of food webs, with evidence indicating this degradation began prior to, and intensified during, the marine extinction interval (Sephton et al., Reference Sephton, Looy, Brinkhuis, Wignall, de Leeuw and Visscher2005; Chu et al., Reference Chu, Grasby, Song, Corso, Wang, Mather, Wu, Song, Shu, Tong and Wignall2020). These processes were accompanied by declines in atmospheric oxygen availability inferred from geochemical model reconstructions (Berner, Reference Berner2009; Mills et al., Reference Mills, Krause, Jarvis and Cramer2023) and by a rapid release of mercury from terrestrial reservoirs (up to ~100-fold above background levels, Dal Corso et al., Reference Dal Corso, Mills, Chu, Newton, Mather, Shu, Wu, Tong and Wignall2020), together exacerbating physiological stress and inducing pervasive hypoxic conditions for terrestrial organisms. The combined hypoxia and hypercapnia compressed the altitudinal ranges available to species, reducing habitat diversity (Huey and Ward, Reference Huey and Ward2005). A significant body of research indicates that the terrestrial faunal collapse was driven by the temperature-dependent metabolic squeeze, where the increased heat necessitated a higher metabolic rate while the available atmospheric oxygen decreased, leading to systemic physiological breakdown (Knoll et al., Reference Knoll, Bambach, Payne, Pruss and Fischer2007; Benton, Reference Benton2018; Penn et al., Reference Penn, Deutsch, Payne and Sperling2018; Chu et al., Reference Chu, Song, Dal Corso, Winguth, Gautam, Wignall, Grasby, Shu, Song, Song, Tian, Wu and Tong2025). Terrestrial species were affected later and less than marine taxa in terms of extinction intensity, with the wave of extinction proceeding from mid-latitudes toward the tropics (Wu et al., Reference Wu, Zhang, Ramezani, Zhang, Erwin, Feng, Shao, Cai, Zhang, Xu and Shen2024). A similar latitudinal pattern characterized the Late Ordovician mass extinction (which was driven by a global glaciation, Finnegan et al., Reference Finnegan, Heim, Peters and Fischer2012). In contrast, the Late Devonian and Late Triassic mass extinctions were more impactful in the tropics (Reddin et al., Reference Reddin, Kocsis and Kiessling2019) although the latter (Davies et al., Reference Davies, Marzoli, Bertrand, Youbi, Ernesto and Schaltegger2017), but not the former (Smart et al., Reference Smart, Filippelli, Gilhooly, Ozaki, Reinhard, Marshall and Whiteside2023), experienced end-Permian like volcanism and global warming. These idiosyncrasies confirm that the feedback mechanisms between climate change, biogeography and biogeochemical cycles and the ensuing changes in extinction risk are nonlinear (Armstrong McKay et al., Reference Armstrong McKay, Staal, Abrams, Winkelmann, Sakschewski, Loriani, Fetzer, Cornell, Rockström and Lenton2022; Kaiho, Reference Kaiho2022; Alsulami and Petrovskii, Reference Alsulami and Petrovskii2023).

The most massive marine Cenozoic extinctions were driven by extreme warming during the early Paleogene hyperthermals (Lourens et al., Reference Lourens, Sluijs, Kroon, Zachos, Thomas, Röhl, Bowles and Raffi2005; Lauretano et al., Reference Lauretano, Littler, Polling, Zachos and Lourens2015). The first, abrupt hyperthermal occurred at the Paleocene Eocene boundary (PETM) (McInerney and Wing, Reference McInerney and Wing2011; Jiang et al., Reference Jiang, Cui, Wang, De Palma, Naafs, Jiang, Hu, Wu, Chu, Gu, Wang, Huang, Ingalls, Bralower, Yang, Zachos and Ridgwell2025), although evidence indicates continental-scale warming preceding its onset (Secord et al., Reference Secord, Gingerich, Lohmann and MacLeod2010). The PETM and the ensuing Eocene Thermal Maximum were characterized by a highly dynamic global warming-dominated climate system, operating across both long (>10⁶ years) and short (<10⁴ years) timescales (Lauretano et al., Reference Lauretano, Littler, Polling, Zachos and Lourens2015). The PETM was short and brutal, lasting less than ~200 kyr with global temperatures rising by ~5–8 °C (Lauretano et al., Reference Lauretano, Littler, Polling, Zachos and Lourens2015). Hyperthermals were initiated by intense volcanism in the North Atlantic (Gutjahr et al., Reference Gutjahr, Ridgwell, Sexton, Anagnostou, Pearson, Pälike, Norris, Thomas and Foster2017) and unfolded within a climate system characterized by El Niño-like variability (Abhik et al., Reference Abhik, Dommenget, McGregor, Hutchinson, Steinig, Zhu, Capitanio, Lunt, Niezgodzki, Knorr, W-L and Abe-Ouchi2025). These conditions promoted widespread ocean deoxygenation and acidification, which in turn drove a vast extinction wave among marine benthic communities, with losses estimated at approximately 30–50% of deep-sea benthic foraminiferal species (Sluijs et al., Reference Sluijs, van Roij, Harrington, Schouten, Sessa, LeVay, Reichart and Slomp2014; Yao et al., Reference Yao, Paytan and Wortmann2018; Haynes and Hönisch, Reference Haynes and Hönisch2020; He et al., Reference He, Kemp, Li and Ruhl2023). On land, hyperthermals forced repeated dwarfing across mammalian lineages (D’Ambrosia et al., Reference D’Ambrosia, Clyde, Fricke, Gingerich and Abels2017), and faunal and floral turnover, but extinction intensity was comparatively modest (Gingerich, Reference Gingerich, Wing, Gingerich, Schmitz and Thomas2003; Wing et al., Reference Wing, Harrington, Smith, Bloch, Boyer and Freeman2005; Bowen et al., Reference Bowen, Bralower, Delaney, Dickens, Kelly, Koch, Kump, Meng, Sloan, Thomas, Wing and Zachos2006; McInerney and Wing, Reference McInerney and Wing2011; Secord et al., Reference Secord, Bloch, Chester, Boyer, Wood, Wing, Kraus, McInerney and Krigbaum2012; Willard et al., Reference Willard, Donders, Reichgelt, Greenwood, Sangiorgi, Peterse, Nierop, Frieling, Schouten and Sluijs2019).

Across the Phanerozoic, several other major extinction events have been linked to large injections of greenhouse gases triggering a recurring suite of kill mechanisms, often referred to as the “deadly trio” (Bijma et al., Reference Bijma, Pörtner, Yesson and Rogers2013) of global warming, ocean acidification and deoxygenation (e.g. end Triassic, Bond et al., Reference Bond, Dickson, Ruhl, Bos and van de Schootbrugge2023), Toarcian, during the Jurassic (Remírez et al., Reference Remírez, Gilleaudeau, Gan, Kipp, Tissot, Kaufman and Parente2024), and Cretaceous (Owens et al., Reference Owens, Gill, Jenkyns, Bates, Severmann, Kuypers, Woodfine and Lyons2013; Petrizzo et al., Reference Petrizzo, Parente, Falzoni, Bottini, Frijia, Steuber and Erba2025). Within this framework, the Permian/Triassic and the Paleocene-Eocene transitions stand out as cases in which rapid global warming played a central role in driving extinction and ecological reorganization at multiple levels (Secord et al., Reference Secord, Bloch, Chester, Boyer, Wood, Wing, Kraus, McInerney and Krigbaum2012; Kiessling et al., Reference Kiessling, Smith and Raja2023; Finnegan et al., Reference Finnegan, Harnik, Lockwood, Lotze, McClenachan and Kahanamoku2024; Hatfield et al., Reference Hatfield, Allen, Carroll, Dean, Deng, Gordon, Guillerme, Hansford, Hoyal Cuthill, Mannion, Martins, Payne, Shipley, Thomas, Thompson, Woods and Davis2025), posing a warning to us as we face another biodiversity crisis (Sperling et al., Reference Sperling, Boag, Duncan, Endriga, Marquez, Mills, Monarrez, Sclafani, Stockey and Payne2022; Finnegan et al., Reference Finnegan, Harnik, Lockwood, Lotze, McClenachan and Kahanamoku2024; Kiessling et al., Reference Kiessling, Reddin, Dowding, Dimitrijević, Raja and Kocsis2025). Although the current warming trend is not as intense in magnitude as during these events, its rapidity is alarmingly close (Olhoff et al., Reference Olhoff, Lamb, Kuramochi, Rogeli, den Elzen, Christensen, Fransen, Pathak and Tong2025). Data from multiple sources demonstrate the link between the magnitude and rate of temperature change and extinction intensity throughout the Phanerozoic (Kaiho et al., Reference Kaiho, Oshima, Adachi, Adachi, Mizukami, Fujibayashi and Saito2016; Song et al., Reference Song, Kemp, Tian, Chu, Song and Dai2021; Kaiho, Reference Kaiho2022). While both cooling and warming events have driven extinctions in the past (Harnik et al., Reference Harnik, Lotze, Anderson, Finkel, Finnegan, Lindberg, Liow, Lockwood, McClain, McGuire, O’Dea, Pandolfi, Simpson and Tittensor2012; Harper et al., Reference Harper, Cascales-Miñana and Servais2020; Morgan et al., Reference Morgan, Bralower, Brugger and Wünnemann2022; Weppe et al., Reference Weppe, Condamine, Guinot, Maugoust and Orliac2023; O’Connor et al., Reference O’Connor, Jerrett, Price, Lyson, Lengger, Peterse and van Dongen2024), the rapid, high-magnitude current warming is hastening chemical and physical changes in the oceans at a rate unprecedented in the paleoclimate record for the last 55 million years (IPCC, 2023) and could reach levels never experienced during the last 300 million years (Caldeira and Wickett, Reference Caldeira and Wickett2003; Hönisch et al., Reference Hönisch, Ridgwell, Schmidt, Thomas, Gibbs, Sluijs, Zeebe, Kump, Martindale, Greene, Kiessling, Ries, Zachos, Royer, Barker, Marchitto, Moyer, Pelejero, Ziveri, Foster and Williams2012). Major differences between current global warming and past events stand in the comparatively lower level of hydrogen sulfide in the atmosphere and less widespread ocean anoxia (Foster et al., Reference Foster, Hull, Lunt and Zachos2018; Penn et al., Reference Penn, Deutsch, Payne and Sperling2018). However, it is important to note that anoxia is not a prerequisite for extinction, as clearly demonstrated by end Triassic global warming whose death toll was immense, but fully anoxic conditions only developed 200–250 kyr after the extinction, during the Hettangian stage (Somlyay et al., Reference Somlyay, Palcsu, Kiss, Clarkson, Kovács, Vallner, Zajzon and Pálfy2023). Still, although anoxia is usually associated with global warming because warm surface water holds less dissolved oxygen and favor water stratification and the development of sulfate-reducing bacteria communities, intense cooling is consistent with some major anoxic events in the recent past (Chen et al., Reference Chen, Montañez, Zhang, Isson, Macarewich, Planavsky, Zhang, Rauzi, Daviau, Yao, Qi, Wang, Fan, Poulsen, Anbar, Shen and Wang2022) and may have contributed to triggering the Late Ordovician mass extinction (Bartlett et al., Reference Bartlett, Elrick, Wheeley, Polyak, Desrochers and Asmerom2018).

Overall, although multiple ocean anoxic events (OAEs) were associated with extreme marine extinction rates (Reershemius and Planavsky, Reference Reershemius and Planavsky2021) terrestrial ecosystems remained comparatively resilient. This contrast underscores a key point: the mechanisms and consequences of global warming differ fundamentally between the marine and terrestrial realms, making it necessary to examine continental systems separately in the sections that follow.

Mammals, climate thresholds and human impacts through the Cenozoic

Cenozoic mammalian responses illustrate the strong context-dependence of extinction and faunal turnover under climate forcing that is typical of continental biodiversity (Kiessling et al., Reference Kiessling, Smith and Raja2023). In contrast with the marine realm, where warming, acidification and deoxygenation acted as potent kill mechanisms, mammals in the Cenozoic terrestrial faunas show more heterogeneous, regionally modulated responses to climate change. This heterogeneity reflects the influence of local ecological structure, dispersal barriers, species interactions and asynchronous evolutionary histories.

Cenozoic climates progressively became cooler on average (Zachos et al., Reference Zachos, Pagani, Sloan, Thomas and Billups2001). The cooling trend accelerated twice. The first episode occurred at the Eocene/Oligocene transition (~34 Ma), while the second acceleration occurred with the onset of fully Ice Age conditions in the Northern Hemisphere at the beginning of the Pleistocene (2.58 Ma) or possibly slightly before (Haug et al., Reference Haug, Ganopolski, Sigman, Rosell-Mele, Swann, Tiedemann, Jaccard, Bollmann, Maslin, Leng and Eglinton2005; Bailey et al., Reference Bailey, Hole, Foster, Wilson, Storey, Trueman and Raymo2013). At that time, the amplitude and intensity of climate change became more extreme, the quality of the fossil and paleoenvironmental record reached its highest (Timmermann et al., Reference Timmermann, Raia, Mondanaro, Zollikofer, Ponce de León, Zeller and Yun2024) and the new biggest extinction factor, the genus Homo, entered the scene.

The Eocene/Oligocene cold spell was sufficiently intense, and its effects long-lasting enough to mark the transition from the greenhouse Earth to the icehouse Earth (Katz et al., Reference Katz, Miller, Wright, Wade, Browning, Cramer and Rosenthal2008). On land, this climatic deterioration was associated with a profound environmental reorganization and correlates in Europe with a major faunal turnover, known as the Grande Coupure. The Coupure is best described as a massive, diachronic turnover event (Prothero, Reference Prothero1994; Hooker et al., Reference Hooker, Collinson and Sille2004), which nonetheless involved sizeable extinction among mammals (de Vries et al., Reference de Vries, Heritage, Borths, Sallam and Seiffert2021; Weppe et al., Reference Weppe, Condamine, Guinot, Maugoust and Orliac2023), crocodylians (Mannion et al., Reference Mannion, Benson, Carrano, Tennant, Judd and Butler2015) and plants (Jaramillo et al., Reference Jaramillo, Rueda and Mora2006). While this turnover coincided with global climatic deterioration, its expression was strongly modulated by regional biogeographic contexts, and the same global climatic transition had little detectable impact on mammalian diversity in North America, underscoring the nonuniform response of terrestrial faunas to global climate change.

The diversity pattern of Cenozoic North American mammals has been suggested to covary with long-term climatic trends (Figueirido et al., Reference Figueirido, Janis, Pérez-Claros, De Renzi and Palmqvist2012). Such covariation has been interpreted as consistent with the idea that large-scale faunal dynamics and the termination of stable mammalian chronofaunas may be subjected to climatic forcing over vast macroevolutionary timescales. Yet, alternative explanations for the pace of biodiversity change exist (Alroy, Reference Alroy2008; Bush and Payne, Reference Bush and Payne2021), and the very essence of mammalian chronofaunas must be considered in the light of diachronous appearance events (Prothero et al., Reference Prothero, Walsh and Alroy1998; Lister and Sher, Reference Lister and Sher2001; Raia et al., Reference Raia, Carotenuto, Meloro, Piras, Barbera and Kotsakis2009). Still, the faunal compositional shifts can be mediated by ecological and biogeographic mechanisms rather than direct climatic forcing (Fortelius et al., Reference Fortelius, Eronen, Kaya, Tang, Raia and Puolamäki2014).

In Europe, the Vallesian Age of the Late Miocene (lasting from 11.1 to 8.7 Ma), witnessed a profound reorganization of mammalian communities, transitioning from the relatively stable, humid and forest-dominated environments of the Middle Miocene to the more seasonal, open and arid landscapes characteristic of the succeeding Turolian Age. While the Early Vallesian is characterized as a period of ecological mixing, where newly arrived immigrants like the equid Hipparion coexisted with relic Middle Miocene forest-dwelling forms such as the suid Listriodon and the cervid Euprox, the onset of higher seasonality, decreased precipitation, and intense cooling at around 9.7 Ma (Domingo et al., Reference Domingo, Koch, Hernández Fernández, Fox, Domingo and Alberdi2013) triggered an extinction wave affecting forest-dwelling taxa and precipitating a major faunal turnover across the continent (Agusti and Moya-Sola, Reference Agusti, Moya-Sola, Kauffman and Walliser1990). Although climate change seems to be the driver, the restructuring of Vallesian ecosystems is best understood as a combination of climatic deterioration, vegetation shifts, dispersal events and the diachronous nature of mammalian chronofaunas (Casanovas-Vilar et al., Reference Casanovas-Vilar, Van den Hoek Ostende, Furió and Madern2014). The same pattern appears for different reasons during the Great American Biotic Interchange (GABI). Both the early Vallesian and GABI produced highly diverse, supersaturated communities where novel immigrants coexisted with endemic faunas for some periods (Webb, Reference Webb2006; Agustí et al., Reference Agustí, Cabrera and Garcés2013), until severe extinctions decimated native species, especially in South America during GABI (Agusti and Moya-Sola, Reference Agusti, Moya-Sola, Kauffman and Walliser1990; Cione et al., Reference Cione, Gasparini, Soibelzon, Soibelzon and Tonni2015), with the contribution of climate change, the immigration of novel placental predators and human hunting (Cione et al., Reference Cione, Tonni and Soilbenzon2003; Faurby and Svenning, Reference Faurby and Svenning2016). In general terms, species-level analyses suggest that global climate change is more consistently associated with increases in turnover and evolutionary rates than with large, synchronous changes in absolute diversity or extinction intensity across continents (Alroy, Reference Alroy2000).

The beginning of the Pleistocene was marked by a similar faunal turnover, although of lower extinction intensity. Unlike the Grande Coupure, which involved substantial extinction, early Pleistocene turnover was dominated by range shifts and taxonomic replacement rather than high extinction intensity. Extinctions driven by the Pliocene/Pleistocene cooling (Brigham-Grette et al., Reference Brigham-Grette, Melles, Minyuk, Andreev, Tarasov, DeConto, Koenig, Nowaczyk, Wennrich, Rosén, Haltia, Cook, Gebhardt, Meyer-Jacob, Snyder and Herzschuh2013) are recorded in Africa (Bibi and Cantalapiedra, Reference Bibi and Cantalapiedra2023), Western Eurasia (Raia et al., Reference Raia, Carotenuto, Meloro, Piras, Barbera and Kotsakis2009) and among marine mammals and other large oceanic vertebrates (Pimiento et al., Reference Pimiento, Griffin, Clements, Silvestro, Varela, Uhen and Jaramillo2017). The intensity of extinction during both events was lower than at the Cretaceous/Paleogene transition, and lower than the current rate (Hatfield et al., Reference Hatfield, Allen, Carroll, Dean, Deng, Gordon, Guillerme, Hansford, Hoyal Cuthill, Mannion, Martins, Payne, Shipley, Thomas, Thompson, Woods and Davis2025). However, the pace of species loss was set to change completely at the end of the Quaternary. Approximately between 50,000 and 5,000 years ago, numerous large-bodied mammals, or the megafauna, disappeared across the globe (Brook and Bowman, Reference Brook and Bowman2002; Koch and Barnosky, Reference Koch and Barnosky2006). This period has been the subject of extensive scientific debate concerning the relative roles of climate change versus human impact as primary drivers, and fueled by the limited availability of direct extinction date estimates (Brook and Bowman, Reference Brook and Bowman2002; Koch and Barnosky, Reference Koch and Barnosky2006; Stuart and Lister, Reference Stuart and Lister2007; Bradshaw et al., Reference Bradshaw, Johnson, Llewelyn, Weisbecker, Strona and Saltré2021).

The role of environmental change in late Quaternary megafauna extinction is multifaceted, with evidence supporting its influence on species vulnerability and availability of suitable habitats. Extinctions often coincided with major climatic shifts, including repeated Heinrich events and other abrupt stadial phases, the Last Glacial Maximum (LGM), and the subsequent warming during the Pleistocene–Holocene transition (Lima-Ribeiro and Diniz-Filho, Reference Lima-Ribeiro and Diniz-Filho2014; Cooper et al., Reference Cooper, Turney, Hughen, Brook, McDonald and Bradshaw2015; Seersholm et al., Reference Seersholm, Werndly, Grealy, Johnson, Keenan Early, Lundelius, Winsborough, Farr, Toomey, Hansen, Shapiro, Waters, McDonald, Linderholm, Stafford and Bunce2020; Graipel et al., Reference Graipel, Lima-Ribeiro, Cherem, Momsen Miller, Diniz-Filho and Cáceres2024). In Eurasia, the LGM created strong selection pressure that led to the disappearance of warm-adapted species, including Palaeoloxodon antiquus and Hippopotamus amphibius (Stuart and Lister, Reference Stuart and Lister2007; Carotenuto et al., Reference Carotenuto, Di Febbraro, Melchionna, Castiglione, Saggese, Serio, Mondanaro, Passaro, Loy and Raia2016). The loss of suitable habitats due to shifting climates was a key mechanism. In South America, the replacement of cold grasslands with Nothofagus forests during warming periods correlated with the local extinction of open-habitat grazers such as Hippidion and Lama (Metcalf et al., Reference Metcalf, Turney, Barnett, Martin, Bray, Vilstrup, Orlando, Salas-Gismondi, Loponte, Medina, De Nigris, Civalero, Fernández, Gasco, Duran, Seymour, Otaola, Gil, Paunero, Prevosti, Bradshaw, Wheeler, Borrero, Austin and Cooper2016). Similarly, in the Great Lakes region of North America, the decline of open spruce parklands during the Younger Dryas likely contributed to the extirpation of mastodons and mammoths (Meltzer, Reference Meltzer2020). Species responded to these changes in different ways. Many species shifted their geographic ranges to track suitable habitats, with those moving over longer distances in Eurasia tending to survive longer (Raia et al., Reference Raia, Passaro, Fulgione and Carotenuto2012). Other populations retreated to areas with less harsh conditions, known as refugia, such as southern Europe (Sommer and Nadachowski, Reference Sommer and Nadachowski2006; Stewart et al., Reference Stewart, Lister, Barnes and Dalén2010). The hypothesis of climate as the sole driver is clearly insufficient to explain end-Pleistocene extinction patterns, especially where humans acted as immigrants and met species naive to their presence (Milberg and Tyrberg, Reference Milberg and Tyrberg1993; Gittleman and Gompper, Reference Gittleman and Gompper2001). Extinctions were not globally synchronous (Surovell et al., Reference Surovell, Pelton, Anderson-Sprecher and Myers2016) but often followed the arrival of humans on different continents, still weakening the idea of a single, global climatic trigger (Brook and Bowman, Reference Brook and Bowman2005; Johnson et al., Reference Johnson, Bradshaw, Cooper, Gillespie and Brook2013). The ‘overkill hypothesis’ (Martin and Leakey, Reference Martin and Leakey1967; Martin and Klein, Reference Martin, Klein, Martin and Klein1984) posits that human hunting was in fact the primary driver of extinction, a view supported by strong chronological coincidence and selectivity, and by explicit modeling of human prey harvesting and population growth (Alroy, Reference Alroy2001; Fordham et al., Reference Fordham, Brown, Akçakaya, Brook, Haythorne, Manica, Shoemaker, Austin, Blonder, Pilowsky, Rahbek and Nogues-Bravo2022, Reference Fordham, Brown, Canteri, Austin, Lomolino, Haythorne, Armstrong, Bocherens, Manica, Rey-Iglesia, Rahbek, Nogués-Bravo and Lorenzen2024; Todorov and Alroy, Reference Todorov and Alroy2025). For some species, synergistic effects of climate change and human overkill are reasonably supported (Nogués-Bravo et al., Reference Nogués-Bravo, Rodríguez, Hortal, Batra and Araújo2008; Wroe et al., Reference Wroe, Field, Archer, Grayson, Price, Louys, Faith, Webb, Davidson and Mooney2013; Mondanaro et al., Reference Mondanaro, Di Febbraro, Melchionna, Carotenuto, Castiglione, Serio, Danisi, Rook, JAFJAF and Raia2019; Meltzer, Reference Meltzer2020), and it is plausible that the temporal sequence of climatic instability followed by human arrival repeatedly created extinction-prone conditions.

Regardless of the extinction driver, the loss of the megafauna caused profound and lasting ecological consequences, leading to significant ecological state shifts. As megaherbivores, such as mammoths and mastodons, ceased acting as ecosystem engineers – a role attached to elephants in modern ecosystems (Owen-Smith, Reference Owen-Smith1988; Tambling et al., Reference Tambling, Minnie, Adendorff and Kerley2013) – landscapes saw an increase in woody vegetation (Owen-Smith, Reference Owen-Smith2001) enhanced fire regimes (Gill, Reference Gill2014), and altered predator to prey ratios, which in turn impacted the survival of smaller species via apparent competition (Meloro et al., Reference Meloro, Raia and Barbera2007; Raia et al., Reference Raia, Meloro and Barbera2007; Mondanaro et al., Reference Mondanaro, Castiglione, Melchionna, Di Febbraro, Vitagliano, Serio, Vero, Carotenuto and Raia2017). Through their activities, the megafauna provided key ecosystem services, including regulation of vegetation structure, maintenance of habitat heterogeneity, and long-distance nutrient redistribution (Doughty et al., Reference Doughty, Wolf and Malhi2013, Reference Doughty, Faurby and Svenning2016; Pedersen et al., Reference Pedersen, Faurby and Svenning2023). By consuming large quantities of plant biomass and dispersing large-seeded fruits over long distances, megafauna facilitated both seed dispersal and gene flow while transporting nutrients across landscapes via dung, urine and carcasses (Malhi et al., Reference Malhi, Doughty, Galetti, Smith, J-C and Terborgh2016). The loss of these processes reduces ecosystem productivity, stability and resistance to environmental perturbations (Smith et al., Reference Smith, Elliott and Lyons2010; Smith et al., Reference Smith, Hammond, Balk, Elliott, Lyons, Pardi, Tomé, Wagner and Westover2016a; Pedersen et al., Reference Pedersen, Faurby and Svenning2023; Carter and Alroy, Reference Carter and Alroy2025), resulting not only in taxonomic impoverishment but also in reduced functionality, and limiting the capacity of ecosystems to buffer climatic variability and recover from disturbance (Malhi et al., Reference Malhi, Doughty, Galetti, Smith, J-C and Terborgh2016; Smith et al., Reference Smith, Tomé, Elliott Smith, Lyons, Newsome and Stafford2016b; Galetti et al., Reference Galetti, Moleón, Jordano, Pires, Guimarães, Pape, Nichols, Hansen, Olesen, Munk, de Mattos, Schweiger, Owen-Smith, Johnson, Marquis and Svenning2018; Tóth et al., Reference Tóth, Lyons, Barr, Behrensmeyer, Blois, Bobe, Davis, Du, Eronen, Faith, Fraser, Gotelli, Graves, Jukar, Miller, Pineda-Munoz, Soul, Villaseñor and Alroy2019; Berti and Svenning, Reference Berti and Svenning2020; Fricke et al., Reference Fricke, Hsieh, Middleton, Gorczynski, Cappello, Sanisidro, Rowan, Svenning and Beaudrot2022; Smith et al., Reference Smith, Elliott Smith, Villaseñor, Tomé, Lyons and Newsome2022; Brook et al., Reference Brook, Lyons, Carter, Gearty, Todorov, Aandahl and Alroy2025). These state shifts, defined as the persistent, nonlinear reorganizations of ecosystem structure and function that result in alternative stable or quasi-stable states following major disturbances (Scheffer et al., Reference Scheffer, Carpenter, Foley, Folke and Walker2001; Barnosky et al., Reference Barnosky, Lindsey, Villavicencio, Bostelmann, Hadly, Wanket and Marshall2016), imply increased extinction risk and further biodivesrsity loss because of biotic homogenization (the increased similarity of species composition among ecological communities (McKinney and Lockwood, Reference McKinney and Lockwood1999; Fraser et al., Reference Fraser, Villaseñor, Tóth, Balk, Eronen, Andrew Barr, Behrensmeyer, Davis, Du, Tyler Faith, Graves, Gotelli, Jukar, Looy, BJ, Miller, Pineda-Munoz, Potts, Shupinski, Soul and Kathleen Lyons2022, but see Keck et al., Reference Keck, Peller, Alther, Barouillet, Blackman, Capo, Chonova, Couton, Fehlinger, Kirschner, Knüsel, Muneret, Oester, Tapolczai, Zhang and Altermatt2025).

Besides making kills, late Pleistocene humans were themselves victims of climate change. Abrupt climatic cooling may have forced extinction in several Homo species, through reduced climatic niche space, fragmentation and exacerbating interspecific competition (Shea, Reference Shea2008; Melchionna et al., Reference Melchionna, Di Febbraro, Carotenuto, Rook, Mondanaro, Castiglione, Serio, Vero, Tesone, Piccolo, Diniz-Filho and Raia2018; Raia et al., Reference Raia, Mondanaro, Melchionna, Di Febbraro, Diniz-Filho, Rangel, Holden, Carotenuto, Edwards, Lima-Ribeiro, Profico, Maiorano, Castiglione, Serio and Rook2020; Hu et al., Reference Hu, Hao, Du, Di Vincenzo, Manzi, Cui, Fu, Pan and Li2023; Margari et al., Reference Margari, Hodell, Parfitt, Ashton, Grimalt, Kim, Yun, Gibbard, Stringer, Timmermann and Tzedakis2023; Muttoni and Kent, Reference Muttoni and Kent2024; ÓhÉigeartaigh, Reference ÓhÉigeartaigh2025). Even Homo sapiens, at once the most successful species and the only survivor of the Homo lineage (Gilpin et al., Reference Gilpin, Feldman and Aoki2016; Zeller et al., Reference Zeller, Timmermann, Yun, Raia, Stein and Ruan2023) almost vanished during the climatic changes triggered by the Toba eruption (Ambrose, Reference Ambrose1998).

The accelerating crisis: Paleoclimate analogies and projections

The evidence from the fossil record examined here, spanning the last 300 million years, from the early Permian through the Cenozoic and Pleistocene, provides reference information for evaluating the severe risk inherent in the current global warming trend. The first and most important lesson to learn is that since the modern biodiversity crisis is driven by direct anthropogenic pressures fragmenting, converting and reducing the extent of natural habitats and their associated biodiversity, the effective temperature increase needed to elicit a mass extinction is probably lower than inferred from deep-time global warming analogs, especially on lands (Figure 1). The initial goal of the 2015 Paris Agreement, limiting warming to 1.5 °C above pre-industrial levels, is now largely understood to be a milestone that the world is almost certain to miss (Tanaka and O’Neill, Reference Tanaka and O’Neill2018; Brecha et al., Reference Brecha, Ganti, Lamboll, Nicholls, Hare, Lewis, Meinshausen, Schaeffer, Smith and Gidden2022; Lee et al., Reference Lee, Calvin, Dasgupta, Krinner, Mukherji, Thorne, Trisos, Romero, Aldunce, Barret, Lee and Romero2023). Recent scientific analyses found that even if every nation fully implements its stated 2030 climate action plans, global warming is projected to stabilize between 2.3 °C and 2.5 °C by the end of the century, and possibly up to 2.8 °C (Olhoff et al., Reference Olhoff, Lamb, Kuramochi, Rogeli, den Elzen, Christensen, Fransen, Pathak and Tong2025). This means that we are off-track with the Paris agreement, and, soberly, more on track toward a scenario consistent with thermal stress thresholds identified for mass extinctions. Deep-time analyses suggest that the warming threshold to trigger a mass extinction is 5.2 °C (Song et al., Reference Song, Kemp, Tian, Chu, Song and Dai2021; Kaiho, Reference Kaiho2022), which is higher than the present-day figures but refers to a world without human-induced disturbance on biodiversity. Anthropogenic activities likely reduce the threshold by bringing species closer to their physiological limits with little space and little time to move towards more favorable habitats. As a matter of fact, early signs of hyperthermal conditions resembling the mass extinctions’ records are already being registered. Current measurements confirm that ocean systems are experiencing fast chemical and physical changes in line with the worst hyperthermals registered over the past 300 Myr (Caldeira and Wickett, Reference Caldeira and Wickett2003; Hönisch et al., Reference Hönisch, Ridgwell, Schmidt, Thomas, Gibbs, Sluijs, Zeebe, Kump, Martindale, Greene, Kiessling, Ries, Zachos, Royer, Barker, Marchitto, Moyer, Pelejero, Ziveri, Foster and Williams2012). The ocean has absorbed approximately 25–30% of anthropogenic carbon dioxide, leading to a drop in average surface pH from 8.2 to approximately 8.1 (Lee et al., Reference Lee, Calvin, Dasgupta, Krinner, Mukherji, Thorne, Trisos, Romero, Aldunce, Barret, Lee and Romero2023). On the pH scale, this represents a nearly 30% increase in acidity, a level that inhibits the ability of calcifying organisms to build their shells (Peters et al., Reference Peters, O’Leary, Hawkins and Roberts2015; Penn and Deutsch, Reference Penn and Deutsch2022) and is happening at a faster rate than during the PETM (Bijma et al., Reference Bijma, Pörtner, Yesson and Rogers2013; Haynes and Hönisch, Reference Haynes and Hönisch2020). These conditions closely resemble end-Permian and end-Triassic environments (Greene et al., Reference Greene, Martindale, Ritterbush, Bottjer, Corsetti and Berelson2012; Clarkson et al., Reference Clarkson, Kasemann, Wood, Lenton, Daines, Richoz, Ohnemueller, Meixner, Poulton and Tipper2015) and were a factor in reef species biodiversity drops over the last 500 Myr (Kiessling and Simpson, Reference Kiessling and Simpson2011). Concurrently, the oceans hold less dissolved oxygen. Coupled with nutrient runoff leading to microbial respiration, this is driving widespread deoxygenation (Stramma et al., Reference Stramma, Johnson, Sprintall and Mohrholz2008; Keeling et al., Reference Keeling, Körtzinger and Gruber2010) to the extent that the global oxygen inventory reduced at least 2% since the mid-20th century (Schmidtko et al., Reference Schmidtko, Stramma and Visbeck2017). The deoxygenation is linked to high atmospheric CO₂ levels today as it was in past (Oschlies et al., Reference Oschlies, Brandt, Stramma and Schmidtko2018; Bauer et al., Reference Bauer, McKenzie, Cheung, Gambacorta, Bottini, Nordsvan, Erba and Crowe2024). The expanding hypoxic ‘dead zones’ (Diaz and Rosenberg, Reference Diaz and Rosenberg2008) particularly noted in coastal areas such as the Gulf of Mexico and the Baltic Sea (Rabalais et al., Reference Rabalais, Turner, Díaz and Justić2009), mirror equally hypoxic seas associated with major marine extinctions (Falkowski et al., Reference Falkowski, Algeo, Codispoti, Deutsch, Emerson, Hales, Huey, Jenkins, Kump, Levin, Lyons, Nelson, Schofield, Summons, Talley, Thomas, Whitney and Pilcher2011). The recent increased uptake of dissolved nitrogen by microorganisms (Sun et al., Reference Sun, Buchanan, Zhang, San Roman, Babbin and Zakem2024a; Beman, Reference Beman2025), is itself a recurrent signature of massive extinction in the oceans, as these microorganisms deplete the oceanic nitrate, reducing ecosystem productivity and increasing toxic ammonium. The same pattern occurred during mass extinction events: end Ordovician (Finney et al., Reference Finney, Berry and Cooper2007), end Devonian (Liu et al., Reference Liu, Qie, Algeo, Yao, Huang and Luo2016) and end Permian (Knies et al., Reference Knies, Grasby, Beauchamp and Schubert2013; Sun et al., Reference Sun, Zulla, Joachimski, Bond, Wignall, Zhang and Zhang2019). The combination of a warmer, more acidic, and less oxygenated ocean constitutes the most significant threat to marine biodiversity, effectively placing the oceans’ ecosystems under chemical and physiological siege (Penn and Deutsch, Reference Penn and Deutsch2022). Finally, although the global relevance of this phenomenon is uncertain, oxidation of continental shelf clathrates can in theory transform the methane they sequester into CO₂, consuming dissolved oxygen (Dickens et al., Reference Dickens, O’Neil, Rea and Owen1995). This positive feedback mechanism can rapidly accelerate ocean’s oxygen depletion and increase warming and acidification with as little as 2 °C increase in sea surface temperature (SST, Babila et al., Reference Babila, Penman, Standish, Doubrawa, Bralower, Robinson, Self-Trail, Speijer, Stassen, Foster and Zachos2022). This figure is equivalent to ~3 °C warming of Earth’s atmosphere, much less than Song et al.’s 5.2 °C temperature threshold.

Figure 1.

Conceptual representation of extinction dynamics under global warming. Schematic illustration of how extinction risk may respond to increasing global temperature in marine and terrestrial realms. The trajectories are conceptual and intended to highlight qualitative differences in system behavior with progressing global warming. Solid trajectories reflect the extinction magnitude versus warming distribution consistent with past hyperthermals event. The dashed lines represent the hypothetical trajectories tweaked by anthropogenic effects, both on lands and at the sea.

In the oceans, extinction risk is shown to increase steeply with warming, reflecting the combined effects of temperature rise, deoxygenation, and acidification (the “deadly trio”) which is buffered by the ocean thermal behavior under modest warming. On land, extinction dynamics under hyperthermals are less intense than in the oceans, and more context-dependent, with risk distinctively increasing in the Anthropocene as warming interacts with human-generated stressors such as habitat loss, fragmentation and direct exploitation.

The position of key temperature thresholds along the horizontal axis is based on published estimates but should be interpreted as indicative ranges rather than precise tipping points. The figure assumes relatively rapid warming trajectories comparable in their ecological consequences to hyperthermal events, implying slower rates of warming would likely lead to different outcomes. Citations for the key temperature thresholds mentioned in the figure, from left to right, refer to the 2025 Emission Gas Report (Olhoff et al., Reference Olhoff, Lamb, Kuramochi, Rogeli, den Elzen, Christensen, Fransen, Pathak and Tong2025), the clathrate gun hypothesis, identifying at 2 °C SST (Babila et al., Reference Babila, Penman, Standish, Doubrawa, Bralower, Robinson, Self-Trail, Speijer, Stassen, Foster and Zachos2022) the critical temperature increase triggering further warming via a feedback loop (roughly corresponding to 3 °C on the continents), the finding that under the worst-case (RCP 8.5) global warming trajectory, ~16.7% of species are threatened with extinction (Urban, Reference Urban2015) and the average warming level (5.2 °C) triggering a mass extinction in the past (Song et al., Reference Song, Kemp, Tian, Chu, Song and Dai2021).

Right: The matrix compares the data coverage of major macroecological and paleontological databases for mammals, meant to help conservation paleobiology efforts to mitigate global warming effects. From top to bottom: EutherianCoP (Mondanaro et al., Reference Mondanaro, Girardi, Castiglione, Timmermann, Zeller, Venugopal, Serio, Melchionna, Esposito, Di Febbraro and Raia2025), PHYLACINE (Faurby et al., Reference Faurby, Davis, Pedersen, Schowanek, Antonelli and Svenning2018), COMBINE (Soria et al., Reference Soria, Pacifici, Di Marco, Stephen and Rondinini2021), Pantheria (Jones et al., Reference Jones, Bielby, Cardillo, Fritz, O’Dell, Orme, Safi, Sechrest, Boakes, Carbone, Connolly, Cutts, Foster, Grenyer, Habib, Plaster, Price, Rigby, Rist, Teacher, Bininda-Emonds, Gittleman, Mace and Purvis2009), NOW (https://nowdatabase.org/, 2026), PaleoDB (https://paleobiodb.org, 2026), PANGAEA (Felden et al., Reference Felden, Möller, Schindler, Huber, Schumacher, Koppe, Diepenbroek and Glöckner2023), the Ecological Register (ecoregister.org, 2026) and Neotoma (Williams et al., Reference Williams, Grimm, Blois, Charles, Davis, Goring, Graham, Smith, Anderson, Arroyo-Cabrales, Ashworth, Betancourt, Bills, Booth, Buckland, Curry, Giesecke, Jackson, Latorre, Nichols, Purdum, Roth, Stryker and Takahara2018).

Figure 1. Long description

Left Panel Line Graph The horizontal axis represents global temperature increase over the -normal- baseline in degrees Celsius. The vertical axis represents extinction magnitude as a percentage of species going extinct. Blue lines represent the marine realm and green lines represent the terrestrial realm. Solid lines illustrate baseline trajectories from deep-time hyperthermals. Dashed lines illustrate idealized trajectories modified by anthropogenic influences. Vertical gridlines mark several temperature benchmarks. Right Panel Database Matrix The right panel contains a table listing nine databases vertically in rows. Data availability is indicated by icons, where a bone icon represents occurrences.

Although current estimates of extinction rates in the marine realm are much lower than on land, the difference vanishes as data shortage is accounted for (Pimm et al., Reference Pimm, Jenkins, Abell, Brooks, Gittleman, Joppa, Raven, Roberts and Sexton2014; Webb and Mindel, Reference Webb and Mindel2015), implying defaunation is ongoing at comparable paces in both environments (McCauley et al., Reference McCauley, Pinsky, Palumbi, Estes, Joyce and Warner2015). Future predictions are equally gloomy since marine species naturally lack the terrestrial species’ opportunity to track their preferred environments or settle within refugia, given their lower thermal tolerance and the mobility of marine habitats (Sunday et al., Reference Sunday, Bates and Dulvy2011; Hodapp et al., Reference Hodapp, Roca, Fiorentino, Garilao, Kaschner, Kesner-Reyes, Schneider, Segschneider, Kocsis, Kiessling, Brey and Froese2023; Albano et al., Reference Albano, Schultz, Wessely, Taviani, Dullinger and Danise2024).

On land, things are even more worrisome. Terrestrial wild species currently constitute 1.8% of the total biomass of mammals (the remaining part being humans plus the livestock). Marine mammals do not fare much better at 3.6% (Greenspoon et al., Reference Greenspoon, Krieger, Sender, Rosenberg, Bar-On, Moran, Antman, Meiri, Roll, Noor and Milo2023). Furthermore, 40% of the wild mammal biomass is concentrated in just 10 species, and most of the mammalian wildlife is losing population size and lives scattered in fragmented habitats, facing heightened extinction risks (Fahrig, Reference Fahrig2003; Haddad et al., Reference Haddad, Brudvig, Clobert, Davies, Gonzalez, Holt, Lovejoy, Sexton, Austin, Collins, Cook, Damschen, Ewers, Foster, Jenkins, King, Laurance, Levey, Margules, Melbourne, Nicholls, Orrock, D-X and Townshend2015; Lino et al., Reference Lino, Fonseca, Rojas, Fischer and Ramos Pereira2019; Mondanaro et al., Reference Mondanaro, Di Febbraro, Melchionna, Maiorano, Di Marco, Edwards, Holden, Castiglione, Rook and Raia2021; Britnell et al., Reference Britnell, Zhu, Kerley and Shultz2023). The signs of climate warming-induced deterioration on extant mammal species are already evident, including miniaturization (Sheridan and Bickford, Reference Sheridan and Bickford2011; Smith et al., Reference Smith, Smith, Lyons and Payne2018) as during the Eocene hyperthermals (Secord et al., Reference Secord, Bloch, Chester, Boyer, Wood, Wing, Kraus, McInerney and Krigbaum2012; D’Ambrosia et al., Reference D’Ambrosia, Clyde, Fricke, Gingerich and Abels2017), range shifts and contraction (Sunday et al., Reference Sunday, Bates and Dulvy2012; Mitchell et al., Reference Mitchell, Snelling, Hetem, Maloney, Strauss and Fuller2018; Beyer and Manica, Reference Beyer and Manica2020), and physiological evidence of thermal stress (Hetem et al., Reference Hetem, Fuller, Maloney and Mitchell2014; Mitchell et al., Reference Mitchell, Snelling, Hetem, Maloney, Strauss and Fuller2018; de la Fuente et al., Reference de la Fuente, Briscoe, Kearney, Williams, Youngentob, Marsh, Cernusak, Leahy, Larson and Krockenberger2025), which is not even sparing the livestock (Nardone et al., Reference Nardone, Ronchi, Lacetera and Bernabucci2006; Thornton et al., Reference Thornton, Nelson, Mayberry and Herrero2021). As with the Late Pleistocene, large mammals are especially targeted because of their vulnerability to warming (Ripple et al., Reference Ripple, Newsome, Wolf, Dirzo, Everatt, Galetti, Hayward, Kerley, Levi, Lindsey, Macdonald, Malhi, Painter, Sandom, Terborgh and Van Valkenburgh2015, Reference Ripple, Wolf, Newsome, Betts, Ceballos, Courchamp, Hayward, Van Valkenburgh, Wallach and Worm2019). Conversely, small mammals, with their high surface-area-to-volume ratio, can exchange heat with the environment more efficiently, granting them a survival advantage under thermal stress (Brook and Bowman, Reference Brook and Bowman2005).

Although most past warming events generally caused lower extinction rates on land than in the oceans, the current trend is, by reference, ten times faster than during PETM (Zeebe et al., Reference Zeebe, Ridgwell and Zachos2016; Gingerich, Reference Gingerich2019; Haynes and Hönisch, Reference Haynes and Hönisch2020), which means 1 °C in 50 years. Such a rapid pace exceeds the adaptive capacity of most terrestrial species (Opdam and Wascher, Reference Opdam and Wascher2004; Burrows et al., Reference Burrows, Schoeman, Buckley, Moore, Poloczanska, Brander, Brown, Bruno, Duarte, Halpern, Holding, Kappel, Kiessling, O’Connor, Pandolfi, Parmesan, Schwing, Sydeman and Richardson2011; Wang et al., Reference Wang, Pineda-Munoz and McGuire2023; Belfiore et al., Reference Belfiore, Mondanaro, Castiglione, Melchionna, Girardi, Raia and Di Febbraro2024), and occurs across human-encroached landscapes, severely limiting the ability of terrestrial organisms to track suitable climates through range shifts (Hof et al., Reference Hof, Araújo, Jetz and Rahbek2011; Sandel et al., Reference Sandel, Arge, Dalsgaard, Davies, Gaston, Sutherland and Svenning2011; Haddad et al., Reference Haddad, Brudvig, Clobert, Davies, Gonzalez, Holt, Lovejoy, Sexton, Austin, Collins, Cook, Damschen, Ewers, Foster, Jenkins, King, Laurance, Levey, Margules, Melbourne, Nicholls, Orrock, D-X and Townshend2015; Urban, Reference Urban2015; Lino et al., Reference Lino, Fonseca, Rojas, Fischer and Ramos Pereira2019; Lenoir et al., Reference Lenoir, Bertrand, Comte, Bourgeaud, Hattab, Murienne and Grenouillet2020). These constraints highlight the need for modern climate-change adaptation strategies grounded in physiological limits and dispersal barriers (McGuire et al., Reference McGuire, Lawler, McRae, Nuñez and Theobald2016; Beger et al., Reference Beger, Metaxas, Balbar, McGowan, Daigle, Kuempel, Treml and Possingham2022). These peculiar, unprecedented settings make the current biodiversity crisis unique and may well imply terrestrial species are in danger more than ever before during the Cenozoic (Hatfield et al., Reference Hatfield, Davis and Thomas2022). This implies that rather than end-Permian or PETM, the current global warming is potentially more reminiscent of the K-Pg or end-Triassic mass extinctions, when the terrestrial fauna was severely hit, at least as much as, if not more than, ocean life (Cribb et al., Reference Cribb, Formoso, Woolley, Beech, Brophy, Byrne, Cassady, Godbold, Larina, Maxeiner, Wu, Corsetti and Bottjer2023). While determining whether or not we are amidst the sixth mass extinction is a scientifically sound and important question (Barnosky et al., Reference Barnosky, Matzke, Tomiya, Wogan, Swartz, Quental, Marshall, McGuire, Lindsey, Maguire, Mersey and Ferrer2011; Ceballos et al., Reference Ceballos, Ehrlich and Dirzo2017; Wiens and Saban, Reference Wiens and Saban2025), studies accounting for data deficiency in poorly studied groups inform that we may have lost as much as ~10% of Earth’s biodiversity during the last 500 years (Cowie et al., Reference Cowie, Bouchet and Fontaine2022). The potentially devastating economic and health consequences of this biodiversity loss call for immediate, unwavering action (Lenton et al., Reference Lenton, Xu, Abrams, Ghadiali, Loriani, Sakschewski, Zimm, Ebi, Dunn, Svenning and Scheffer2023; Pörtner et al., Reference Pörtner, Scholes, Arneth, Barnes, Burrows, Diamond, Duarte, Kiessling, Leadley, Managi, McElwee, Midgley, Ngo, Obura, Pascual, Sankaran, Shin and Val2023).

The macroeconomics of climate inertia

The economic and existential costs of the global warming trajectory are staggering. Some estimates suggest it will cost as much as 12% Gross Domestic Product (GDP) for each degree of global temperature increase and hundreds of thousands of lives in the next few decades (Kemp et al., Reference Kemp, Xu, Depledge, Ebi, Gibbins, Kohler, Rockström, Scheffer, Schellnhuber, Steffen and Lenton2022; Lenton et al., Reference Lenton, Xu, Abrams, Ghadiali, Loriani, Sakschewski, Zimm, Ebi, Dunn, Svenning and Scheffer2023; Pörtner et al., Reference Pörtner, Scholes, Arneth, Barnes, Burrows, Diamond, Duarte, Kiessling, Leadley, Managi, McElwee, Midgley, Ngo, Obura, Pascual, Sankaran, Shin and Val2023; Bilal Diego Känzig et al., Reference Bilal and Ȩnzig2024; Callahan et al., Reference Callahan, Trok, Wilson, Gould, Heft-Neal, Diffenbaugh and Burke2025). Therefore, the impact of implementing an immediate policy shift overturning the current immobility of political and economic systems – and the fallout if such a change does not occur – is central theme in macroeconomics and political economy. This inertia is not an accident, but a consequence of market and governmental systems driven by self-interest. The delay in implementing corrective climate policy is fundamentally rooted in short-termism and collective action failure (Basurto and Ostrom, Reference Basurto and Ostrom2009). Political and corporate entities operate on short cycles (elections, quarterly earnings) that prioritize immediate financial stability over the long-term, nonmonetized benefits of climate action. The global nature of climate change creates a powerful incentive for nations and corporations to ‘free-ride’, benefiting from the emissions cuts of others while bearing minimal domestic costs, leading to a collective failure to protect the shared natural resources, a typical ‘tragedy of the commons’ situation (Hardin, Reference Hardin1968). The flaw is then amplified by powerful actors engaging in rent-seeking (Helm, Reference Helm2010), which is known to limit effective policies such as cap-and-trade programs (Rode, Reference Rode2021). Rent-seeking describes the use of political and institutional influence, such as lobbying, to extract a larger share of existing wealth, rather than creating new wealth through innovation. In the context of climate policy, this manifests as the political obstruction and regulatory capture by incumbent fossil fuel interests aiming to protect the value of their vast, carbon-intensive assets against the necessary transition (MacKenzie, Reference MacKenzie2017). This persistent preference for short-term private gain over long-term public interest drives the global economy toward the high-risk, 2.5–2.8 °C trajectory.

The primary failure identified is the continued reliance on economic frameworks (e.g., GDP, standard discount rates) that treat nature as a substitutable factor of production and biodiversity loss as an incremental, linear cost (Raworth, Reference Raworth2018; Dasgupta, Reference Dasgupta2021; Lean, Reference Lean2025). Given that humanity is rapidly eroding its natural asset base, it is increasingly necessary to recognize the natural capital as a fundamental, nonsubstitutable asset, internalizing the long-term, extinction-level costs of climate change that are currently externalized by short-term rent-seeking behaviors, thus preserving intergenerational justice (Browne, Reference Browne2025). Economic discounting devalues future environmental costs, which is incompatible with the temporal scale of extinction. Contemporary ecological governance is transitioning from simplistic reforestation targets, following the scientific recalibration of mass-planting initiatives (Bastin et al., Reference Bastin, Finegold, Garcia, Mollicone, Rezende, Routh, Zohner and Crowther2019; Fleischman et al., Reference Fleischman, Basant, Chhatre, Coleman, Fischer, Gupta, Güneralp, Kashwan, Khatri, Muscarella, Powers, Ramprasad, Rana, Solorzano and Veldman2020), toward performance-based finance models like the Tropical Forest Forever Facility (Tropical Forest Forever Facility Secretariat, 2025). Concurrently, the 30 × 30 marine mandate is advancing through the legal framework of the High Seas Treaty, though effective protection remains hindered by substantial financing gaps (For the Ocean, 2025). Ultimately, stabilizing planetary boundaries requires internalizing environmental externalities through True Cost Accounting (Food and of the United Nations, 2023) and addressing the ethical imperative for Western nations to reduce unsustainable material consumption (United Nations Environment Programme, 2024). These and other solutions presented in the literature are increasingly synergistic (Pörtner et al., Reference Pörtner, Scholes, Arneth, Barnes, Burrows, Diamond, Duarte, Kiessling, Leadley, Managi, McElwee, Midgley, Ngo, Obura, Pascual, Sankaran, Shin and Val2023) and point to the need to protect biodiversity by implementing nature-based solutions for climate mitigation that take into account the ‘natural’ cost of resources (Roe et al., Reference Roe, Streck, Beach, Busch, Chapman, Daioglou, Deppermann, Doelman, Emmet-Booth, Engelmann, Fricko, Frischmann, Funk, Grassi, Griscom, Havlik, Hanssen, Humpenöder, Landholm, Lomax, Lehmann, Mesnildrey, Nabuurs, Popp, Rivard, Sanderman, Sohngen, Smith, Stehfest, Woolf and Lawrence2021).