Rethinking megafauna

1. Introduction

Prehistoric art provides evidence that megafauna (literally, ‘large animals’; see electronic supplementary material, appendix S1 for the etymology and popular definitions of this term) have fascinated humans since our origins (e.g. [1]). The eminent nineteenth-century naturalist Wallace [2] referred to megafauna as ‘the hugest, and fiercest, and strangest forms'. A hundred and forty plus years later, however, megafaunal research still lacks a unifying framework for the use of this term, which has diverged in the development of disciplines as diverse as wildlife biology, oceanography, limnology, soil ecology, evolutionary biology, conservation biology, palaeontology and anthropology. Thus, definitions in the scientific literature include disparate combinations of species: from the smallest organisms readily visible in photographs to the largest vertebrates ever on earth (e.g. [3–5]; figure 1, electronic supplementary material, appendix S2). Given the great sociocultural significance of megafauna [6,7], the ubiquity of the megafauna concept in addressing profound and varied scientific questions [8–11], and the multiple threats that jeopardize large animals [12–14], a re-examination of the concept is warranted [15].

• Download figure
• Open in new tab
• Download PowerPoint

Here, we review the concept of megafauna and propose a goal-oriented framework for megafauna research, which may support scientific endeavours, improve conservation policy and practice, and strengthen the public perception. To do this, we adopt a two-pronged approach. First, we review the scientific literature to (i) examine the different definitions of megafauna and (ii) analyse the terminology commonly associated with the concept of megafauna. Second, we carry out a survey among ecologists and palaeontologists to (iii) assess the traits of the species they consider as megafauna and (iv) identify the key criteria that should define megafauna. The goal of this survey is to enhance our understanding of how researchers working with megafauna conceptualize data that already exist in the scientific literature. Based on insights gained from the review and survey, we propose a working scheme for the use of the megafauna concept, discuss pros and cons of different definitions, and provide recommendations for advancing interdisciplinary megafaunal research.

2. Literature review

(a) Megafauna definitions

We conducted a systematic review of existing megafauna definitions in the scientific literature (276 articles reviewed; see electronic supplementary material, appendix S3 for a complete list of references and electronic supplementary material, appendix S4 for the searching methods). The majority of megafauna articles focused on terrestrial species (55% of the papers; mainly concerned with prehistorical times) and marine ecosystems (52%; mostly referencing recent times), with very few articles dealing with freshwater megafauna (1%; figure 2 and electronic supplementary material, figure S1). Our search did not uncover any paper dealing with soil megafauna, although soil ecologists use this term as well [16].

• Download figure
• Open in new tab
• Download PowerPoint

When considering whether the reviewed papers provided definitions of the term megafauna and how such definitions were justified, strikingly, 74% of the identified articles did not provide an explicit definition of megafauna. Among the remaining 26% (i.e. the 71 articles using a definition), 45% did not provide any argument or reference to support the definition, whereas 25% provided references, 20% specified distinct arguments and 10% offered both references and arguments (figure 2). Definitions, when provided, were somewhat idiosyncratic (i.e. varied according to the study system) and relied on ad hoc size-related criteria (see electronic supplementary material, table S1 and figure 1; for a complete list of definitions, see electronic supplementary material, table S2).

Definitions of the megafauna concept were primarily of two types. The first group used an explicit, albeit generally arbitrary, body-size threshold above which a species is considered megafauna. Among the definitions of this group, a distinction can be made between those that used a mass-based threshold and those that used a length-based threshold.

On the one hand, mass thresholds ranging from around 10 kg to 2 tons have been widely used in a terrestrial context to define megafauna [5]. Palaeontologists, for example, have often referred to the megafauna definition provided by Martin [4]: i.e. animals, usually mammals, over 100 pounds (ca 45 kg; e.g. [17–20]). Recently, this megafauna definition has also been applied to marine environments [21], and several authors have adopted a slightly lower threshold (30 kg) to define freshwater megafauna [14,22]. Some terrestrial megafauna studies (e.g. [23]) are based on the megaherbivore concept of Owen-Smith [24,25], restricted to herbivores exceeding 1000 kg in adult body mass according to distinctions from smaller herbivores in a number of ecological features. Other authors have applied guild-dependent thresholds for terrestrial megafauna (e.g. greater than or equal to 100 kg for herbivores and greater than or equal to 15 kg for carnivores) [13]. Finally, Hansen and Galetti [26] emphasized the importance of taking into account the ecological context too: ‘one ecosystem's mesofauna is another ecosystem's megafauna’. This means that relatively small species can also be considered megafauna, as long as they are, or were, among the largest species occurring in a given area.

On the other hand, papers in which the megafauna definition relies on body length are characterized by much smaller size thresholds. These studies have been common in the context of benthic and epibenthic environments, where marine megafauna are usually defined as animals visible on seabed photographs (normally over ca 1 cm) or caught by trawl nets (e.g. [3,27–29]). Furthermore, soil ecologists have used the term megafauna to encompass those species above 20 mm in length that exert strong influences on gross soil structure [16].

The second major group of papers included those that relied on body size only implicitly—i.e. considering megafauna as certain clades or groups of species that are relatively large-sized within the focal study system. These articles normally concerned aquatic environments. Several studies of marine benthic megafauna focused on particular taxonomic groups, such as decapods and fish [30,31]. In a marine pelagic context, some authors focused on the largest sea-dwelling species—i.e. marine mammals, sea turtles and seabirds (termed ‘air-breathing marine megafauna’) [32], along with sharks, rays and other predatory fish (e.g. [33–35]) and even polar bears and cephalopods [36]. In freshwater ecosystems, crustaceans, amphibians and fish were classified as megafauna by some authors [37]. Other work has focused on particular functional groups, such as higher/apex marine predators [34,36]. It is noteworthy that the term megafauna has been virtually ignored for dinosaurs and, until recently, barely used for mammals other than those of the Late Pleistocene period. Instead, dinosaur experts and wildlife biologists prefer using the species, clade or group name rather than the more general term megafauna (e.g. [38–41]).

(b) Terminology associated with megafauna research

As demonstrated above, the megafauna definition may differ according to the studied ecosystem. In this section, we highlight the fact that definitions also differ depending on the ecological and biological questions of the study. To this end, we created semantic networks based on the terms included in the title and abstract of the 276 reviewed articles, and identified thematic clusters based on co-occurrence of these terms (see electronic supplementary material, appendix S4 for methodological details). From this, we obtained three major megafauna research clusters (electronic supplementary material, figures S1 and S2). The first cluster included articles on terrestrial megafauna and mainly corresponded to the study of the extinction of Pleistocene megafauna: its timing, causes and impacts on ecosystems (e.g. [17,42,43]). The terms included in this terrestrial cluster were related to the megafauna definitions provided by Owen-Smith [24] and, mostly, by Martin [4]. The second cluster concerned extant benthic and epibenthic marine megafauna: the characterization of their communities [44–46], the environmental factors that determine their composition [47–49] and their ecological properties [9,30]. In general, the terms of this cluster were linked to definitions not specifying a body-size threshold [3,32]. The third cluster covered studies on the impacts of bycatch in fisheries, mainly on marine air-breathing vertebrates [12,32,50], as well as on strategies for their conservation [51,52].

These clusters were not totally disconnected, as electronic supplementary material, figure S2 reveals several bridging terms that have the potential to link different clusters in the network [53]. For example, terrestrial and pelagic clusters were recently connected by research on the conservation of threatened vertebrates in relation to global change [54–57]. In this case, important bridging terms were impact, climate and review (electronic supplementary material, figure S2). Similarly, benthic and pelagic clusters were interlinked by research on biodiversity conservation in marine environments [58], with biodiversity, use and fish being bridging terms (electronic supplementary material, figure S2). Thus, our lexical analysis revealed a growing, albeit still weak, tendency to connect the different conceptual clusters that make up the main megafauna research network. Our findings indicate that the increasing concern about the causes and consequences of human impacts on the conservation of large animals has a promising potential to foster collaboration among researchers focusing on different ecosystems (e.g. [59]).

3. Survey of researchers

Given that the majority of the papers using the concept megafauna do not provide a definition of this term, we surveyed researchers working on megafauna to get a better understanding of how they understand the concept when using it.

(a) Species traits associated with megafauna

To understand the species traits (i.e. taxonomy, biology, ecology, behaviour, conservation status and popularity; see electronic supplementary material, tables S3 and S4 for more details) that researchers associated with megafauna, we asked ecologists and palaeontologists (n = 93 respondents) to fill in a questionnaire that included photos of 120 animal species (electronic supplementary material, table S3). In the questionnaire, respondents had to specify which species they considered as megafauna. Then we ranked species traits according to their capacity to predict the probability that the respondents would classify these species as megafauna (see electronic supplementary material, appendix S4 and tables S3–S5 for methodological details). We found that adult body mass was by far the most important trait, followed by the taxonomic group; all other traits analysed were of minor importance (electronic supplementary material, figure S3a). According to a generalized linear model (GLM), body mass and taxonomic group accurately predicted the probability that a species would be classified as megafauna (F_15,104 = 72.79, p < 0.001, R² = 0.90). Larger species were more likely to be considered as megafauna, following a sigmoidal (logistic) relationship (figure 3a). However, the slope of this relationship varied among taxonomic groups, as reflected by the significance of the interaction coefficient (F_7,104 = 4.13, p < 0.001; figure 3b). Mammals, birds and reptiles had steeper slopes, fish species had intermediate values, and amphibians and invertebrates exhibited shallower slopes (figure 3b). Thus, for a given body mass, the classification of a species as megafauna depended on its taxonomy, likely reflecting a bias arising from the prominence of terrestrial vertebrate species in scientific research or the general (average) size of the species in the different groups. These patterns were consistent despite variability in respondents' characteristics such as age and expertize (see electronic supplementary material, appendix S4 and figures S3b and S4).

• Download figure
• Open in new tab
• Download PowerPoint

(b) What criteria should define megafauna?

We also used the questionnaire to assess researchers’ recommendations for defining megafauna. We explicitly asked the respondents to choose among six criteria needed to define megafauna: body mass, taxonomy, ecological function, ecological context, life-history traits and extinction risk. Respondents could choose as many of them as they wanted and could also name additional criteria (see electronic supplementary material, appendix S4 for methodological details). Among the criteria provided, 92% of respondents identified body mass as the key criterion (electronic supplementary material, figure S5). However, body mass was very often (86% of respondents) chosen in combination with other criteria (mean total number ± s.d. of criteria selected by respondents: 2.9 ± 1.3). This suggests that body size alone is insufficient for defining megafauna. Extinction risk was rarely taken into account in defining megafauna, probably because respondents identified this criterion as a circular and extrinsic argument or because it cannot be applied to extinct taxa, which frequently contributed to megafauna research. The selection of criteria was again barely affected by respondents' characteristics (see electronic supplementary material, table S6, figures S6 and S7). Only 7% of the respondents suggested alternative criteria to define megafauna. These additional suggestions (namely species’ volume, habitat requirements, ‘importance’ within the food web, ecological ‘status’, ecosystem and temporal context) were closely related to the six criteria already provided in the questionnaires.

4. Rethinking the megafauna concept

As evidenced in the literature, the term megafauna has been widely applied in ecological and palaeontological research. However, our literature review revealed that researchers have been adopting a context-dependent use of the term, most often using operational definitions with varying and largely arbitrary body-size thresholds and taxonomic groups as proxies, depending on the study system and research question. Only a few studies have explicitly emphasized the functional importance of the largest species in a given ecosystem and over a specific period [16,24,26]. In addition, our survey of researchers provided consensus that body size (e.g. body mass) is a crucial descriptor, but not necessarily sufficient, for addressing the different applications of the term megafauna.

When rethinking the megafauna concept, the primary question that should arise is whether we need a threshold. As argued next, there are reasons that justify the search for non-arbitrary thresholds and that indicate that these are, in fact, achievable, at least in some cases. First, avoiding a threshold-based definition would make the use of the megafauna term largely impractical. Second, clear breakpoints in either body size or ecological features have been identified for some animal groups (see below). Thus, a follow-up agenda exploring whether corresponding thresholds do, or do not exist in different groups of organisms is needed.

Below, we reconsider the megafauna concept and propose a general working scheme for its use in various ecological and evolutionary contexts. These include either natural systems (i.e. before Homo sapiens began to defaunate them [26]) or systems that have been impacted by human-mediated extinctions and introductions of wild and domestic species [60].

(a) The largest

The central challenge in using a threshold concept to define megafauna—as is also the case for other popular ecological terms such as keystone, flagship or umbrella species (see [61])—is how to empirically establish a metric (e.g. body mass, or body length) and a corresponding value above which an animal may be effectively regarded as megafauna. This value needs to be placed within a community or an ecosystem context to make any sense. We could circumvent this threshold concept by simply defining ‘megafauna’ as the subset of largest species in a community or an ecosystem. To answer the critical question of what the threshold should be, we could follow two approaches. In its simplest form, we could refer to the single largest species. Going beyond this, a transparent definition of ‘subset’ requires exploring the frequency distributions of body size (e.g. body mass) values within the community or ecosystem under study, and determining a breakpoint in body size. Although body size data are not available for all animal species within an ecosystem, this information is often biased towards larger species [62].

Another approach would be to focus on particular clades or guilds to restrict the species pool under consideration, facilitating the identification of megafauna. Thus, ‘clade- or guild-specific megafauna’ would be the subset of largest species of a given clade or guild in a community or an ecosystem. This implies acknowledging that the megafauna within a clade or guild do not necessarily include the largest species in the ecosystem. Within phylogenetic lineages, body mass is skewed towards smaller sizes, with larger species being almost invariably rarer than smaller species [24,63,64]. For instance, greater than 90% of sub-Saharan vertebrate herbivore species weigh less than 500 kg, while only ca 5% of species has a body mass exceeding 1000 kg [24]. However, most animals, with the exceptions of birds and mammals, grow through prolonged ontogenetic stages. For instance, giant bluefin tuna (Thunnus thynnus) covers 5–6 orders of magnitude in mass from larvae to adult [65]. Whether scales of ontogenetic change cause taxa with long developmental changes in size to have a shallower slope than in cases where the break might be more obvious needs to be investigated.

(b) Operational definitions

We refer to operational definitions as those using specific body size criteria but that are not based on a body size distribution, namely most definitions enumerated in the electronic supplementary material, tables S1 and S2. A prominent example is Martin's definition of megafauna (ca 45 kg [4]), which can be seen as a human-centred perspective, partitioning animals similar or larger in size than humans from those smaller. These definitions have been the core of the megafauna scientific literature, most likely because of their obvious practical advantages. For instance, they facilitate data processing and analysis, and they may normally apply to both extant and extinct species.

The main feature of operational definitions is their strong dependence on the research discipline, which makes them highly applicable to conduct comparisons within disciplines but strongly limits their trans-disciplinary use. However, some attempts have recently been made to move certain operational definitions beyond the original research context. In particular, the application or adaptation of Martin's megafauna standard [4] to aquatic environments [14,21,22] represents a connection among terrestrial, marine pelagic and freshwater megafauna research. In addition, soil and marine benthos megafauna research, which is concerned with communities characterized by relatively small-sized species, may be closely linked because they use similar—body length-based—definitions. However, a weak connection between terrestrial/marine pelagic/freshwater and soil/benthos megafauna research is anticipated due to their very different conceptions of ‘mega’ (figure 1). Nevertheless, while operational definitions could seem conducive to multidisciplinary coordination and collaboration in megafauna research (e.g. to undertake biodiversity inventories and conservation status assessments), the application of operational thresholds to different disciplines relies on the unrealistic assumption that body mass (and functional traits; see below) distributions are comparable among different communities or ecosystems. Thus, operational definitions, which are inherently arbitrary, are at risk of including or ignoring species that respectively should or should not be considered as megafauna, in both intra- and cross-disciplinary approaches.

(c) Functional definitions: looking for a new approach

While some existing definitions go beyond body size (e.g. [16,26]), we largely lack a conceptual definition of megafauna that integrates the ecological function and functional traits of a species along with its size (e.g. represented by body mass; but see [24]; figure 4). In this section, we present a function-oriented framework for the use of the megafauna concept, therefore, responding to the general perception of researchers that body size alone is an incomplete descriptor of megafauna (see above). Here, unlike previous definitions, which were primarily based on body size, breakpoints are associated with biological and ecological features/qualities that vary with body size. These functional concepts can be applied to different communities and ecosystems, from terrestrial and soil to marine and freshwater systems, and are, at least a priori, not biased towards vertebrates or invertebrates.

• Download figure
• Open in new tab
• Download PowerPoint

The first concept, which combines a body-size based megafauna definition with the keystone species concept [69], assumes that the largest species in an ecosystem generally have disproportionally large effects on the structure and functioning of their communities and ecosystems, both in magnitude and in the spatial and temporal heterogeneity they create [70]. In line with this concept, a disproportionate increase in energy use (e.g. represented by population biomass) in relation to body mass increases has been identified in many vertebrate [24,63] and invertebrate phylogenetic groups [64]. Accordingly, ‘keystone megafauna’ would be the subset of animals among the largest in size that have consistently strong effects on the structure or functioning of a community or an ecosystem. Smaller animals would exhibit high variation in relation to the effects that they exert on their ecosystems, from very weak to very strong (figure 4a). All species that have a strong influence on their ecosystems, in general, stronger than expected by their abundance or biomass, may be regarded as keystone species [61,66–69], but only those with relatively large body size should be termed as keystone megafauna (figure 4b). In practice, this concept of megafauna may require extensive ecological knowledge of the biotic communities and their functioning [66], which would encourage a research agenda to better understand the ecological roles of large species [61,66]. However, the use of proxies for ecological effects, such as size-density relationships [63], could greatly simplify the identification of keystone megafauna within different clades or guilds, including extinct fauna. Comparing the magnitude, variability and skewness, as well as related breakpoints, of these relationships (figure 4a for a general formulation) among different animal groups seems an exciting avenue for future megafauna research.

The second functional concept for megafauna is referred to as ‘functional megafauna’, which can be defined as the subset of largest species of a given clade or guild that have distinctive functional traits (sensu [71]). An important practical advantage of this concept is that the identification of megafauna could be relatively easily accomplished because it only needs a basic ecological knowledge. Ideally, studies should focus on traits with high interspecific variation, that may be easily measurable and, therefore, comparable among the members of a given animal group. For instance, within terrestrial mammals, megaherbivores differ from smaller herbivores in almost all ecological and life-history aspects (e.g. age at first conception, birth interval and gestation time [24]). Also in terrestrial mammals, there is a functional transition associated with a number of life-history traits between carnivores exceeding an average mass of 13–16 kg and those carnivores of smaller size [72]. In other, less-studied cases, the key question is, of course, to define the subset of functional traits to be explored.

A feasible variant of the functional megafauna concept would be ‘apex megafauna’: animals so large that they have escaped most non-anthropogenic predation as adults. This concept is related to the megaherbivore and apex predator concepts [24,25,72] and can be applied to humans too. In Africa, herbivores larger than 150 kg are subject to reduced predation rates than smaller mammalian prey in some areas [73], but only for herbivores exceeding 1000 kg predation is a consistently negligible cause of adult mortality [24,73,74]. Within the order Carnivora, an average mass of ca 15 kg corresponds to the transition between extrinsic- and self-regulation [72].

5. Conclusion

Our comprehensive literature review and survey of researchers point to a dichotomy between the need to establish operational body-size thresholds and a more functional definition of megafauna. This confirms that the concept of megafauna is far from simple, and, probably, it should not be simplified either. However, we highlight that assessing megafauna from a functional perspective could challenge the perception that there may not be a unifying definition of megafauna that can be applied to all eco-evolutionary contexts and scientific approaches. The functional framework we present, which arises from the perception of megafauna researchers that body size is insufficient to capture the varied eco-evolutionary ramifications of megafauna, could help to reach ecological generality and to minimize the arbitrariness of operational and other non-functional definitions, which present ambiguity problems even at the within-discipline level. This requires exploring thresholds in ecological functions and functional traits of animals pertaining to different clades, guilds, communities and ecosystems. Addressing this challenge could help to broaden out megafauna research, and provides an opportunity to increase our biological understanding of megafauna too. Interestingly, important advances have already been made in terrestrial mammalian systems, so that herbivores exceeding 1000 kg and carnivores above an average body mass of ca 15 kg could be considered as paradigmatic examples of both functional and apex megafauna. Until studies exploring other animal groups and ecosystems are available, we encourage scientists to define megafauna unambiguously and clearly present the distinct logic behind their definition in every megafaunal study. Only by being explicit and appropriately contextualizing the concept will we be able to reach the needed conceptual disambiguation.

We found that cross-disciplinary investigations of megafauna are virtually non-existent (but see e.g. [59]), which may be due, in part, to the fact that most megafauna definitions in the scientific literature are strongly context-dependent. The existence of recurrent topics among megafauna researchers concerned with different animal taxa and ecosystems, such as the conservation of threatened megafauna, compels the search for unifying tools. Using functional, rather than arbitrary, operational definitions, would facilitate understanding and cooperation among wildlife, evolutionary and conservation biologists, marine and soil ecologists, limnologists and palaeontologists, and eventually promote cutting-edge research across systems, disciplines, and geographical boundaries [75,76].

Data accessibility

Data and code to replicate analyses are available from the Dryad Digital Repository: https://dx.doi.org/10.5061/dryad.dv41ns1v3 [77].

Authors' contributions

M.M., J.A.S.-Z., J.A.D., E.R. and K.T. conceived and designed the study; M.M. undertook the literature review and collected data; M.M. and Z.M.-R. created the databases; M.M., C.G.-C. and B.M.-L. conducted the semantic and statistical analyses, with critical inputs from all co-authors; M.M. drafted the manuscript; all authors participated in discussions, contributed critically to data interpretation and manuscript reviewing and gave final approval for publication.

Competing interests

We declare we have no competing interests.

Funding

This article was inspired by the workshop ‘Megafauna: from human–wildlife conflicts to ecosystem services' held in 2016 and jointly funded by the Leibniz-Institute of Freshwater Ecology and Inland Fisheries (IGB-Berlin, Germany) and the Estación Biológica de Doñana (EBD-CSIC, Spain). M.M. acknowledges financial support through the Severo Ochoa Program for Centres of Excellence in R+D+I (SEV-2012-0262) and by a research contract Ramón y Cajal from the MINECO (RYC-2015-19231). C.G.-C. is supported by a ‘Juan de la Cierva’ research contract (MINECO, FJCI-2015-25785), and Z.M.-R. by a postdoctoral contract co-funded by the Generalitat Valenciana and the European Social Fund (APOSTD/2019/016). M.G. thanks to Programa BIOTA from Fundação de Amparo à Pesquisa do Estado de São Paulo (2014/01986-0) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). F.H. acknowledges the Erasmus Mundus Joint Doctorate Program SMART (Science for MAnagement of Rivers and their Tidal systems) funded by the European Union. N.S. was supported by the National Science Center in Poland (DEC-2013/08/M/NZ9/00469). J.-C.S. considers this work a contribution to his Carlsberg Foundation Semper Ardens project MegaPast2Future (CF16-0005) and to his VILLUM Investigator project (VILLUM FONDEN, grant 16549). S.C.J. acknowledges funding from the German Federal Ministry of Education and Research (BMBF) for ‘GLANCE’ (Global change effects in river ecosystems; 01LN1320A) N.G. acknowledges financial support through the project PGC2018-093925-B-C33.

Footnotes

Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.4860798.