PNAS PLUS

Biological annihilation via the ongoing sixth mass

extinction signaled by vertebrate population
losses and declines
Gerardo Ceballosa,1, Paul R. Ehrlichb,1, and Rodolfo Dirzob
a
Instituto de Ecología, Universidad Nacional Autónoma de México, Mexico City 04510, Mexico; and bDepartment of Biology, Stanford University, Stanford,
CA 94305

Contributed by Paul R. Ehrlich, May 23, 2017 (sent for review March 28, 2017; reviewed by Thomas E. Lovejoy and Peter H. Raven)

The population extinction pulse we describe here shows, from a especially because many of those species were obscure and had
quantitative viewpoint, that Earth’s sixth mass extinction is more limited ranges, such as the Catarina pupfish (Megupsilon aporus,
severe than perceived when looking exclusively at species extinc- extinct in 2014), a tiny fish from Mexico, or the Christmas Island
tions. Therefore, humanity needs to address anthropogenic popula- pipistrelle (Pipistrellus murrayi, extinct in 2009), a bat that van-
tion extirpation and decimation immediately. That conclusion is ished from its namesake volcanic remnant.
based on analyses of the numbers and degrees of range contraction Species extinctions are obviously very important in the long run,
(indicative of population shrinkage and/or population extinctions because such losses are irreversible and may have profound effects
according to the International Union for Conservation of Nature) ranging from the depletion of Earth’s inspirational and esthetic
using a sample of 27,600 vertebrate species, and on a more detailed resources to deterioration of ecosystem function and services (e.g.,
analysis documenting the population extinctions between 1900 and refs. 17–20). The strong focus among scientists on species extinc-
2015 in 177 mammal species. We find that the rate of population loss tions, however, conveys a common impression that Earth’s biota is
in terrestrial vertebrates is extremely high—even in “species of low not dramatically threatened, or is just slowly entering an episode of
concern.” In our sample, comprising nearly half of known vertebrate major biodiversity loss that need not generate deep concern now
species, 32% (8,851/27,600) are decreasing; that is, they have de- (e.g., ref. 21, but see also refs. 9, 11, 22). Thus, there might be
creased in population size and range. In the 177 mammals for which sufficient time to address the decay of biodiversity later, or to
we have detailed data, all have lost 30% or more of their geographic develop technologies for “deextinction”—the possibility of the
ranges and more than 40% of the species have experienced severe latter being an especially dangerous misimpression (see ref. 23).
population declines (>80% range shrinkage). Our data indicate that Specifically, this approach has led to the neglect of two critical
beyond global species extinctions Earth is experiencing a huge epi- aspects of the present extinction episode: (i) the disappearance of
sode of population declines and extirpations, which will have nega-
populations, which essentially always precedes species extinctions,
tive cascading consequences on ecosystem functioning and services
and (ii) the rapid decrease in numbers of individuals within some
vital to sustaining civilization. We describe this as a “biological an-
of the remaining populations. A detailed analysis of the loss of
nihilation” to highlight the current magnitude of Earth’s ongoing
individuals and populations makes the problem much clearer and
sixth major extinction event.
more worrisome, and highlights a whole set of parameters that are
increasingly critical in considering the Anthropocene’s biological
|
sixth mass extinction population declines | population extinctions | extinction crisis.
|
conservation ecosystem service

Significance
T he loss of biological diversity is one of the most severe human-
caused global environmental problems. Hundreds of species
and myriad populations are being driven to extinction every year The strong focus on species extinctions, a critical aspect of the
(1–8). From the perspective of geological time, Earth’s richest biota contemporary pulse of biological extinction, leads to a common
misimpression that Earth’s biota is not immediately threatened,
ever is already well into a sixth mass extinction episode (9–14).
just slowly entering an episode of major biodiversity loss. This
Mass extinction episodes detected in the fossil record have been

ECOLOGY
view overlooks the current trends of population declines and
measured in terms of rates of global extinctions of species or higher
extinctions. Using a sample of 27,600 terrestrial vertebrate spe-
taxa (e.g., ref. 9). For example, conservatively almost 200 species of
cies, and a more detailed analysis of 177 mammal species, we
vertebrates have gone extinct in the last 100 y. These represent the
show the extremely high degree of population decay in verte-
loss of about 2 species per year. Few realize, however, that if
brates, even in common “species of low concern.” Dwindling
subjected to the estimated “background” or “normal” extinction
population sizes and range shrinkages amount to a massive
rate prevailing in the last 2 million years, the 200 vertebrate species
anthropogenic erosion of biodiversity and of the ecosystem
losses would have taken not a century, but up to 10,000 y to dis-
services essential to civilization. This “biological annihilation”
appear, depending on the animal group analyzed (11). Considering
underlines the seriousness for humanity of Earth’s ongoing sixth
the marine realm, specifically, only 15 animal species have been
mass extinction event.
recorded as globally extinct (15), likely an underestimate, given the
difficulty of accurately recording marine extinctions. Regarding Author contributions: G.C., P.R.E., and R.D. designed research; G.C. and P.R.E. performed
global extinction of invertebrates, available information is limited research; G.C., P.R.E., and R.D. contributed new reagents/analytic tools; G.C. analyzed
and largely focused on threat level. For example, it is estimated data; and G.C., P.R.E., and R.D. wrote the paper.
that 42% of 3,623 terrestrial invertebrate species, and 25% of Reviewers: T.E.L., George Mason University; and P.H.R., Missouri Botanical Garden.
1,306 species of marine invertebrates assessed on the International The authors declare no conflict of interest.
Union for Conservation of Nature (IUCN) Red List are classified Freely available online through the PNAS open access option.
as threatened with extinction (16). However, from the perspective 1
To whom correspondence may be addressed. Email: [email protected] or
Downloaded by guest on September 20, 2021

of a human lifetime it is difficult to appreciate the current mag- [email protected].

nitude of species extinctions. A rate of two vertebrate species ex- This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
tinctions per year does not generate enough public concern, 1073/pnas.1704949114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1704949114 PNAS | Published online July 10, 2017 | E6089–E6096

 In the last few decades, habitat loss, overexploitation, invasive species, particularly strong in mammals and birds, which largely
organisms, pollution, toxification, and more recently climate disrup- drive the overall land vertebrate pattern (Fig. 3, Center). Notably,
tion, as well as the interactions among these factors, have led to the some parts of the planet harbor low absolute numbers of verte-
catastrophic declines in both the numbers and sizes of populations of brate species undergoing decline (Figs. 2 and 3), such as those
both common and rare vertebrate species (24–28). For example, areas of low species richness located in hypercold (northernmost
several species of mammals that were relatively safe one or two locations, particularly of the Western Hemisphere) and hyperarid
decades ago are now endangered. In 2016, there were only (Saharan Africa and Central Asia) regions. However, it is in-
7,000 cheetahs in existence (29) and less than 5,000 Borneo and structive to examine their corresponding proportional numbers, an
Sumatran orangutans (Pongo pygmaeus and P. abelli, respectively) aspect we discuss in detail in another section below.
(28). Populations of African lion (Panthera leo) dropped 43% since The number of decreasing species of all land vertebrates in each
1993 (30), pangolin (Manis spp.) populations have been decimated of the 10,000-km2 quadrats over Earth’s land surface ranges from a
(31), and populations of giraffes dropped from around 115,000 indi- few to more than 365 (Fig. 2). As expected, large concentrations of
viduals thought to be conspecific in 1985, to around 97,000 repre- decreasing vertebrate species occur in species-rich areas of moist
senting what is now recognized to be four species (Giraffa giraffa, G. tropical forests adjacent to mountainous regions, such as the
tippelskirchi, G. reticulata, and G. camelopardalis) in 2015 (32). Andes–Amazon region, the Congo basin-adjacent eastern African
An important antecedent to our work (25) used the number of highlands, and the Himalayas–south Asian jungle belt. The dis-
genetic populations per unit area and then estimated potential loss tribution of the number of decreasing species considering verte-
on the basis of deforestation estimates and the species–area re- brate classes separately reveals notable differences. First, the
lationship (SAR). Given the recognized limitations of the use of maximum number of decreasing species in a 10,000-km2 quadrat
SAR to estimate extinctions, our work provides an approach based varies from a high value of 296 decreasing birds per quadrat, to a
on reduction of species range as a proxy of population extirpation. low maximum of 60 decreasing reptiles in a quadrat. Second,
The most recent Living Planet Index (LPI) has estimated that mammals and birds have relatively similar distribution patterns of
wildlife abundance on the planet decreased by as much as 58%
between 1970 and 2012 (4). The present study is different from LPI
and other related publications in several ways, including that here
we use all decreasing species of vertebrates according to IUCN,
mapping and comparing absolute and relative numbers of species,
and focusing on population losses. Previous estimates seem vali-
dated by the data we present here on the loss of local populations
and the severe decrease in the population size of many others (see
also refs. 3, 4, 6–8, 26). Here we examine the magnitude of losses of
populations of land vertebrate species on a global system of 10,000-km2
quadrats (Methods). Species vary from common to rare, so that
our analysis, which includes all land vertebrate species (am-
phibians, birds, reptiles, and mammals) deemed as “decreasing” by
IUCN, provides a better estimate of population losses than using
exclusively IUCN data on species at risk. Obviously, common spe-
cies decreasing are not ordinarily classified as species at risk. IUCN
criteria provide quantitative thresholds for population size, trend,
and range size, to determine decreasing species (28, 33). We also
evaluate shrinking ranges and population declines for 177 species
of mammals for which data are available on geographic range
shrinkage from ∼1900 to 2015. We specifically focus on local ex-
tinctions by addressing the following questions: (i) What are the
numbers and geographic distributions of decreasing terrestrial ver-
tebrate species (i.e., experiencing population losses)? (ii) What are
the vertebrate groups and geographic regions that have the highest
numbers and proportions of decreasing species? (iii) What is the
scale of local population declines in mammals—a proxy for other
vertebrates? By addressing these questions, we conclude that an-
thropogenic population extinctions amount to a massive erosion of
the greatest biological diversity in the history of Earth and that
population losses and declines are especially important, because it is
populations of organisms that primarily supply the ecosystem ser-
vices so critical to humanity at local and regional levels.
Results
Patterns of Variation in Population Loss Among Vertebrates. Consid-
ering all land vertebrates, our spatially explicit analyses indicate a
massive pulse of population losses, with a global epidemic of
species declines. Those analyses support the view that the decay of
vertebrate animal life is widespread geographically, crosses phy-
logenetic lineages, and involves species ranging in abundance from
Fig. 1. Decreasing land vertebrates, as exemplified with these four species,
common to rare (Figs. 1–4). The losses, however, are not uniform: include taxa with different conservation status (e.g., low concern, critically
some regions exhibit higher concentrations of species with local endangered), current geographic range (e.g., large, very restricted), and
Downloaded by guest on September 20, 2021

population extinctions than others, including a strong latitudinal abundance (e.g., common, rare). The data on conservation status, current
signal corresponding to an intertropical peak (i.e., roughly between geographic range, and abundance are from IUCN (28). Barn swallow image
the Tropics of Cancer and Capricorn) of number of decreasing courtesy of Daniel Garza Galindo (photographer).

E6090 | www.pnas.org/cgi/doi/10.1073/pnas.1704949114 Ceballos et al.

 PNAS PLUS
ECOLOGY

Fig. 2. Global distribution of terrestrial vertebrate species according to IUCN (28). (Left) Global distribution of species richness as indicated by number of species
in each 10,000-km2 quadrat. (Center) Absolute number of decreasing species per quadrat. (Right) Percentage of species that are suffering population losses in
relation to total species richness per quadrat. The maps highlight that regions of known high species richness harbor large absolute numbers of species expe-
riencing high levels of decline and population loss (particularly evident in the Amazon, the central African region, and south/southeast Asia), whereas the
proportion of decreasing species per quadrat shows a strong high-latitude and Saharan Africa signal. In addition, there are several centers of population decline in
both absolute and relative terms (Borneo, for example).

decreasing species, except that birds have more decreasing species and amphibians clearly differ from each other in regions where de-
in the temperate zones. Third, mammals and birds have patterns of creasing species are concentrated. For example, there are more de-
decreasing species quite distinct from those of reptiles and amphib- creasing reptiles in the Eurasian and African continents, and more
Downloaded by guest on September 20, 2021

ians (Figs. 2 and 3), given that the latter are rarer in the northern and decreasing amphibians in the Americas.
southern temperate and subpolar regions (both are essentially absent There is also great variation in the total population size and
from the Arctic and are missing from the Antarctic). Fourth, reptiles geographic ranges among individual species. Although there is no

Ceballos et al. PNAS | Published online July 10, 2017 | E6091

 Fig. 3. Latitudinal distribution of species richness (Left), decreasing species (Center), and the percentage of species (Right) that are suffering population
losses in relation to total species richness, in each 10,000-km2 quadrat. Patterns of species richness in relation to latitude are similar in all vertebrates, although
there are more species per quadrat in birds and mammals and, as expected, a scarcity of reptiles and amphibians at high latitudes. The patterns of number of
species with decreasing populations indicate that regions with high species richness also have high numbers of decreasing species, but the percentage of
decreasing species in relation to species richness shows contrasting patterns between mammals and birds compared with reptiles and amphibians. In
mammals and birds, the percentage of decreasing species is relatively similar in regions with low and high species richness. In contrast, there are propor-
tionally more decreasing species of reptiles and amphibians in regions with low species richness.
Downloaded by guest on September 20, 2021

accurate information on population size for most taxa, whatever is decreasing populations varies from fewer than 100 individuals in
available indicates that the total population size in species with critically endangered species such as the Hainan black-crested

E6092 | www.pnas.org/cgi/doi/10.1073/pnas.1704949114 Ceballos et al.

 PNAS PLUS
Roughly a third (8,851/27,600) of all land vertebrate species
examined are experiencing declines and local population losses of
a considerable magnitude (Figs. 2–4). Such proportion of de-
creasing species varies, depending on the taxonomic group, from
30% or more in the case of mammals, birds, and reptiles, to 15%
in the case of amphibians. Furthermore, of the decreasing species,
many are now considered endangered (Fig. 4). Beyond that,
roughly 30% of all decreasing species are still sufficiently common
that they are considered of “low concern” by IUCN, rather than
“endangered.” That so many common species are decreasing is a
strong sign of the seriousness of the overall contemporary bi-
ological extinction episode.
In our 10,000-km2 quadrats, the proportion of decreasing
species ranges from less than 10% to more than 50% (Fig. 2). The
geographic distributions of absolute (i.e., number) and relative
(i.e., percentage) of decreasing species is contrasting. Whereas
tropical regions have larger numbers of decreasing species, as
expected, given their higher species richness, their corresponding
proportions are relatively low. In contrast, temperate regions tend
Fig. 4. The percentage of decreasing species classified by IUCN as “endangered” to have similar or higher proportions of decreasing species, a trend
(including “critically endangered,” “endangered,” “vulnerable,” and “near- dramatically prominent in the case of reptiles.
threatened”) or “low concern” (including “low concern” and “data-deficient”)
in terrestrial vertebrates. This figure emphasizes that even species that have not Local Population Extinctions in Mammals. Our most detailed data
yet been classified as endangered (roughly 30% in the case of all vertebrates)
allow comparison of historic and present geographic range of a
are declining. This situation is exacerbated in the case of birds, for which close
to 55% of the decreasing species are still classified as “low concern.”
sample of 177 mammal species (Figs. 5 and 6). Most of the
177 mammal species we sampled have lost more than 40% of their
geographic ranges in historic times, and almost half have lost more
gibbon (Nomascus hainanus), to many millions of individuals in than 80% of their ranges in the period ∼1900–2015. At the con-
decreasing common species such as the barn swallow (Hirundo tinental and subcontinental level, some patterns become evident
rustica). Similarly, the smallest ranges (i.e., <1 km2) are seen in (Fig. 5). The predominant category of range contraction is ≥80%
species such as the Carrizal seedeater (Amaurospiza carrizalensis) in Africa (56% of the sampled mammal species), Asia (75% of the
from Venezuela and Herrera’s false coral snake (Lampropeltis species), Australia (60% of the species), and Europe (40% of the
herrerae) from Mexico, both denizens of tiny islands. The largest species). In the Americas, range contractions are less marked but
ranges are hundreds of thousands of square kilometers, as in still considerable: 22% of the species in North America and 17%
the bush dog (Speothos venaticus) from South America and the of the species in South America have experienced range contrac-
common lizard (Zootoca vivipara) from Eurasia. The sum of tions of at least 80%. Nevertheless, 50% of the species in North
the 10,000-km2 quadrats representing the current ranges of the America and 28% of the species in South America have experi-
8,851 decreasing vertebrate species is 1,350,876 quadrats. A highly enced a range contraction of 41% or more.
conservative estimate would indicate a similar number of local The comparison of the 1900–2015 geographic ranges showed
populations facing extinction. This is, of course, a very rough es- that the 177 species of mammals have disappeared from 58,000
timate of the total number of populations, as the number of grid cells. On the assumption that on average each of the 10,000-km2
populations of a decreasing species in each quadrat largely de- occupied quadrats held a single population of the species found
pends, aside from suitable habitat distribution within the quadrat, within it, this implies that roughly 58,000 populations of the
on animal body mass and trophic position (e.g., ref. 34). The as- 177 mammals we examined have gone extinct. Consider the
following emblematic cases: The lion (Panthera leo) was historically
sumption of one population per 10,000 km2 might seem very con-
distributed over most of Africa, southern Europe, and the Middle
servative, as this area could accommodate many populations of
small animals (e.g., 0.1-kg rodents), most of which could have been

ECOLOGY
extirpated. However, 10,000 km2 may not be sufficient for, or can
barely accommodate a viable population of large carnivores (say a
330-kg Siberian tiger; ref. 34). Nonetheless, our results provide
evidence of the extremely large numbers of vertebrate populations
facing extinction, compared with the number of species.

Proportion of Vertebrate Species Decreasing. The proportion of

decreasing vertebrates shows that there are areas across the planet
with high concentrations of decreasing species in all vertebrates
and regions with high proportions of decreasing species of a par-
ticular group (Figs. 2, 3, and 5). For example, in mammals, the
highest percentage of decreasing species is concentrated in tropical
regions, mostly in the Neotropics and Southeast Asia, whereas in
reptiles, the proportional decline concentrates almost exclusively in
Madagascar. Decreasing amphibians are prominent in Mexico, Fig. 5. The percentage of species of land mammals from five major conti-
nents/subcontinents and the entire globe undergoing different degrees (in
Central America, the northern Andes, and Brazil’s Atlantic forest
percentage) of decline in the period ∼1900–2015. Considering the sampled
in the Americas; West Africa and Madagascar in Africa; and India species globally, 56% of them have lost more than 60% of their range, a
Downloaded by guest on September 20, 2021

and Southeast Asia, including Indonesia and Philippines in Asia– pattern that is generally consistent in Africa, Asia, Australia, and Europe,
Southeast Asia. Finally, decreasing species of birds are found over whereas in South America and North America, 35–40% of the species have
large regions of all continents (Fig. 2). experienced range contractions of only 20% or less. (See text for details.)

Ceballos et al. PNAS | Published online July 10, 2017 | E6093

 Discussion
It has recently been shown, using conservative estimates of current
and background species extinction rates, that Earth is now in a
period of mass global species extinction for vertebrate animals
(11). But the true extent of this mass extinction has been under-
estimated, because of the emphasis on species extinction. This
underestimate largely traces to overlooking the accelerating ex-
tinction of populations. Whereas scientists have known for a long
time that several relatively well-studied species have undergone
major contraction of their ranges, experienced considerable pop-
ulation decreases, and suffered many population extinctions, the
global extent of population shrinkage and extirpation has pre-
viously not been recognized and quantified.
In addition, some studies document that invertebrates and plants
are suffering massive losses of populations and species (35–38).
Here we extend investigation of mass extinction to terrestrial ver-
tebrate population decreases and losses, and give estimates of the
number of their species with decreasing populations. The accuracy
of the estimates is strongly dependent on an unknown parameter,
namely, the actual average area occupied by a vertebrate pop-
ulation (e.g., refs. 35, 39–41). However, even if a population would,
on average, occupy an area five times larger than what we have
used here (i.e., 50,000 km2) there would still be hundreds of
thousands of populations that have suffered extinction in the past
few centuries. On the other hand, most vertebrates (∼70%) are
small species of mammals, birds, reptiles, and amphibians. If, on
average, they have one population every 10 km2 then vertebrates
would have suffered more than a billion population extinctions.
Our results show that population extinction in land vertebrates
is geographically omnipresent, but with notable prominence in
tropical, species-rich regions. It is interesting, however, that
when population extinctions are evaluated as the percentage of
total species richness, temperate regions, with their typical low
species diversity, show higher proportions of population loss.
There are some illustrative qualitative examples of population
Fig. 6. Percentage of local population extinction in 177 species of mammals
decreases and their consequences within terrestrial and marine
in 1° × 1° quadrats, as an indication of the severity of the mass extinction vertebrates, but ours is an attempt at a quantitative evaluation of
crises. The maps were generated by comparing historic and current geo- global trends in population extinctions. Recent reviews indicate that
graphic ranges (49) (SI Appendix, SI Methods). Note that large regions in all species extinctions, population decreases, and range contraction
continents have lost 50% or more of the populations of the evaluated (implying population extinctions) among terrestrial invertebrates
mammals. Because of the small sample size, biased to large mammal species, and plants are as severe as among vertebrates (e.g., refs. 35–38). For
this figure can only be used to visualize likely trends in population losses. example, long-term monitoring of insect populations in the United
Kingdom shows that 30–60% of species per taxonomic order have
contracting ranges (36). The situation in plants has been less
East, all the way to northwestern India (SI Appendix, Fig. S1). It is
evaluated; thus it is difficult to compare them with animals, but
now confined to scattered populations in sub-Saharan Africa and a there is little reason to believe that the extinction situation in plants
remnant population in the Gir forest of India. The vast majority of is dramatically different (37). Furthermore, research shows that the
lion populations are gone. In its African stronghold, it historically loss of animal populations indirectly leads to changes in plant
occupied roughly two thousand 10,000-km2 cells, and now it is re- communities (20, 37, 39), frequently causing the reduction of local
duced to some 600 cells. Other species, such as the mountain lion species richness and dominance of a few plant taxa that either ex-
(Puma concolor), are known to be doing better. The mountain lion perience “ecological release” in response to decreasing herbivore
has lost some of its local populations in North America, but has not pressures (42, 43), and/or experience population reductions due to
suffered such disastrous losses as its Old World relative, adapting the decline of animals responsible for pollination or dispersal (e.g.,
relatively well to human-dominated landscapes, and it is still found refs. 2−3, 20). The status of biodiversity among microorganisms is
across 85% of its historic range. too poorly known to permit us to make any comparison and gen-
Clearly, the extinction of mammal populations, although varying eralizations about the current pulse of extinctions, although some
from species to species, has been a global phenomenon (Fig. 6). recent research has unraveled feedbacks between local large her-
Strikingly, the predominant color code in the mammalian map is bivore defaunation and mycorrhizal richness (44, 45). Given what
we know about genetic population differentiation, it is expected
that of 70% or more of population losses, with the exception of
that the range contractions and declines we document here imply a
some areas of South America and high latitudes of North America.
considerable loss of intraspecific genetic diversity (23) but this is,
Particularly hard hit have been the mammals of south and south- clearly, an aspect that warrants further investigation.
east Asia, where all of the large-bodied species of mammals ana- In sum, by losing populations (and species) of vertebrates, we are
lyzed have lost more than 80% of their geographic ranges. The losing intricate ecological networks involving animals, plants, and
Downloaded by guest on September 20, 2021

Cape and Sahara regions in Africa, central Australia, the eastern microorganisms (e.g., refs. 2, 8, 18, 45, 46). We are also losing pools
United States, and the Atlantic forest in South America have also of genetic information that may prove vital to species’ evolutionary
suffered severely from population extinctions. adjustment and survival in a rapidly changing global environment.

E6094 | www.pnas.org/cgi/doi/10.1073/pnas.1704949114 Ceballos et al.

 PNAS PLUS
This suggests that, even if there was not ample sign that the crisis (11, 48). All signs point to ever more powerful assaults on bio-
extends far beyond that group of animals, today’s planetary diversity in the next two decades, painting a dismal picture of the
defaunation of vertebrates will itself promote cascading cata- future of life, including human life.
strophic effects on ecosystems, worsening the annihilation of nature
(2, 3, 46). Thus, while the biosphere is undergoing mass species Methods
extinction (11), it is also being ravaged by a much more serious and For full methods, please see SI Appendix. We determined the number of de-
rapid wave of population declines and extinctions. In combination, creasing vertebrate species using the IUCN (28) Red List of Threatened Species.
these assaults are causing a vast reduction of the fauna and flora of In the IUCN, species are classified as decreasing, stable, or increasing (see also
our planet. The resulting biological annihilation obviously will also ref. 33). Either range contraction (population extinction) or reduction in
have serious ecological, economic, and social consequences (46). numbers in extant populations determines whether a species is decreasing. We
Humanity will eventually pay a very high price for the decimation of used the IUCN maps of terrestrial vertebrates (i.e., mammals, birds, reptiles, and
the only assemblage of life that we know of in the universe. amphibians) to create the global maps of number of species (richness) and of
decreasing species, and percentage of decreasing species in relation to total
Conclusion species richness. The distribution of all of the species was superimposed in a
22,000 grid of 10,000-km2 quadrats covering the continental lands. For the
Population extinctions today are orders of magnitude more fre- grid, a Lambert azimuthal equal-area projection was used (see ref. 49 for de-
quent than species extinctions. Population extinctions, however, are tails of the projection methods). In our analyses a critical issue is how grid
a prelude to species extinctions, so Earth’s sixth mass extinction squares and populations correspond. This is a very difficult problem that varies
episode has proceeded further than most assume. The massive loss with definitions of species. (In this paper, we stick with the classic biological
of populations is already damaging the services ecosystems provide definition of species.) The number of populations also varies from species to
to civilization. When considering this frightening assault on the species; for example, a highly phylopatric species would have more populations
foundations of human civilization, one must never forget that per square than a very vagile species, and species with different mating systems
Earth’s capacity to support life, including human life, has been would have different estimates of numbers of Mendelian populations, and
shaped by life itself (47). When public mention is made of the these would not be the same as estimates of number of demographic units
extinction crisis, it usually focuses on a few animal species (hun- (50). For the purposes of understanding the annihilation, these differences are
dreds out of millions) known to have gone extinct, and projecting not critical. For example, if we have lost 90% of the lion’s geographic range,
many more extinctions in the future. But a glance at our maps whether this amounts to 10,000 demographic units or 4,000 Mendelian pop-
ulations is trivial in the present context. It would be extremely useful if we had
presents a much more realistic picture: they suggest that as much as
much more information on population structure for all vertebrates, but this is a
50% of the number of animal individuals that once shared Earth
major, pending agenda.
with us are already gone, as are billions of populations. Further-
The population extinction analysis was conducted on 177 mammalian species
more, our analysis is conservative, given the increasing trajectories occurring on five continents. Specifically, we analyzed 54 species in Africa, 14 in
of the drivers of extinction and their synergistic effects. Future Asia, 57 in Australia, 15 in Europe, and 35 in America. The historical distribution
losses easily may amount to a further rapid defaunation of the was gathered from specialized literature (see details in ref. 26) and the current
globe and comparable losses in the diversity of plants (36), in- distribution from IUCN (28). Historic and current ranges were digitized as
cluding the local (and eventually global) defaunation-driven coex- geographic information system polygons and elaborated in ArcGis 10.1 (51).
tinction of plants (3, 20). The likelihood of this rapid defaunation For each species, we calculated the area of the historical and present distri-
lies in the proximate causes of population extinctions: habitat bution (in square kilometers) to estimate the percentage of lost area and the
conversion, climate disruption, overexploitation, toxification, spe- percentage of area where the species are extant. A caveat of these estimates
cies invasions, disease, and (potentially) large-scale nuclear war— regards how representative the sample of 177 species is. We recognize a bias in
all tied to one another in complex patterns and usually reinforcing that the data include a large number of medium- and large-sized species, for
each other’s impacts. Much less frequently mentioned are, however, which the best information is available. However, given that such medium and
the ultimate drivers of those immediate causes of biotic destruction, large species are the most seriously threatened by the predominant proxi-
namely, human overpopulation and continued population mate drivers of defaunation (2, 3), the likely bias against small-sized spe-
cies should not affect our overall interpretation of results.
growth, and overconsumption, especially by the rich. These drivers,
all of which trace to the fiction that perpetual growth can occur on a
ACKNOWLEDGMENTS. We thank John Harte for very helpful comments on
finite planet, are themselves increasing rapidly. Thus, we emphasize the manuscript and Noé Torres, Giulia Santulli, and Jesús Pacheco for their
that the sixth mass extinction is already here and the window for help with data analyses. The Universidad Nacional Autónoma de México
effective action is very short, probably two or three decades at most and Stanford University supported our work.

ECOLOGY
1. Ehrlich P-R (1995) The scale of the human enterprise and biodiversity loss. Extinction 13. McCallum ML (2015) Vertebrate biodiversity losses point to a sixth mass extinction.
Rates, eds Lawton JH, May RM (Oxford Univ Press, Oxford, UK), pp 214–226. Biol Conserv 24:2497–2519.
2. Dirzo R, et al. (2014) Defaunation in the Anthropocene. Science 345:401–406. 14. Pimm SL, et al. (2014) The biodiversity of species and their rates of extinction, dis-
3. Young HS, McCauley DJ, Galleti M, Dirzo R (2016) Patterns, causes, and consequences tribution, and protection. Science 344:1246752.
of Anthropocene defaunation. Annu Rev Ecol Evol Syst 47:433–458. 15. McCauley DJ, et al. (2015) Marine defaunation: Animal loss in the global ocean.
4. World Wide Fund for Nature (2016) Living Planet Report 2016. Risk and resilience in a Science 347:1255641.
new era. (WWF International, Gland, Switzerland). Available at wwf.panda.org/about_our_ 16. Collen B, Böhm M, Kemp R, Baillie J (2012) Spineless: Status and Trends of the World’s
earth/all_publications/lpr_2016/. Accessed June 10, 2017. Invertebrates (Zoological Society of London, London).
5. Maxwell SL, Fuller RA, Brooks TM, Watson JEM (2016) Biodiversity: The ravages of 17. Daily G (1997) Nature’s Services: Societal Dependence on Natural Ecosystems. (Island
guns, nets and bulldozers. Nature 536:143–145. Press, Covello, CA).
6. Laliberte AS, Ripple WJ (2004) Range contractions of North American carnivores and 18. Naeem S, Duffy JE, Zavaleta E (2012) The functions of biological diversity in an age of
ungulates. BioScience 54:123–138. extinction. Science 336:1401–1406.
7. Worm B, Tittensor DP (2011) Range contraction in large pelagic predators. Proc Natl 19. Estes JA, et al. (2011) Trophic downgrading of planet Earth. Science 333:301–306.
Acad Sci USA 108:11942–11947. 20. Brosi BJ, Briggs HM (2013) Single pollinator species losses reduce floral fidelity and
8. Ripple WJ, et al. (2014) Status and ecological effects of the world’s largest carnivores. plant reproductive function. Proc Natl Acad Sci USA 110:13044–13048.
Science 343:1241484. 21. Briggs JC (2014) Global biodiversity gain is concurrent with decreasing population
9. Barnosky AD, et al. (2011) Has the Earth’s sixth mass extinction already arrived? sizes. Biodiver J 5:447–452.
Nature 471:51–57. 22. Hooper DU, et al. (2012) A global synthesis reveals biodiversity loss as a major driver
10. Ceballos G, García A, Ehrlich PR (2010) The sixth extinction crisis: Loss of animal of ecosystem change. Nature 486:105–108.
populations and species. J. Cosmology 8:1821–1831. 23. Ehrlich PR (2014) The case against de-extinction: It’s a fascinating but dumb idea.
Downloaded by guest on September 20, 2021

11. Ceballos G, et al. (2015) Accelerated modern human-induced species losses: Entering Yale Environment 360 (Yale University, New Haven, CT). Available at bit.ly/1gAIuJF).
the sixth mass extinction. Sci Adv 1:e1400253. Accessed June 10, 2017.
12. Wake DB, Vredenburg VT (2008) Colloquium paper: Are we in the midst of the sixth mass 24. Hobbs RJ, Mooney HA (1998) Broadening the extinction debate: Population deletions
extinction? A view from the world of amphibians. Proc Natl Acad Sci USA 105:11466–11473. and additions in California and Western Australia. Conserv Biol 12:271–283.

Ceballos et al. PNAS | Published online July 10, 2017 | E6095

 25. Hughes JB, Daily GC, Ehrlich PR (1997) Population diversity: Its extent and extinction. 39. Cardinale BJ, et al. (2012) Biodiversity loss and its impact on humanity. Nature 486:
Science 278:689–692. 59–67.
26. Ceballos G, Ehrlich PR (2002) Mammal population losses and the extinction crisis. 40. Hurlbert AH, Jetz W (2007) Species richness, hotspots, and the scale dependence of
Science 296:904–907. range maps in ecology and conservation. Proc Natl Acad Sci USA 104:13384–13389.
27. Gaston KJ, Fuller RA (2008) Commonness, population depletion and conservation 41. Peterson AT, Navarro‐Sigüenza AG, Gordillo A (2016) Assumption- versus data-based
biology. Trends Ecol Evol 23:14–19. approaches to summarizing species’ ranges. Conserv Biol, 10.1111/cobi.12801.
28. International Union of Conservation of Nature (2015) The IUCN Red List of Threat- 42. Martínez-Ramos M, Ortíz-Rodríguez I, Piñero D, Dirzo R, Sarukhán J (2016) Humans
ened Species, Version 2015.2 (IUCN, 2015). Available at www.iucnredlist.org. Accessed disrupt ecological processes within tropical rainforest reserves. Proc Natl Acad Sci USA
February 10, 2016. Revised January 10, 2017. 113:5323–5328.
29. Durant SM, et al. (2017) The global decline of cheetah Acinonyx jubatus and what it 43. Camargo-Sanabria AA, Mendoza E, Guevara R, Martínez-Ramos M, Dirzo R (2015)
Experimental defaunation of terrestrial mammalian herbivores alters tropical rain-
means for conservation. Proc Natl Acad Sci USA 114:528–533.
forest understorey diversity. Proc Biol Sci 282:20142580.
30. Henschel P, et al. (2014) The lion in West Africa is critically endangered. PLoS One
44. Petipas RH, Brody AK (2014) Termites and ungulates affect arbuscular mycorrhizal
9:e83500.
richness and infectivity in a semiarid savanna. Botany 92:233–240.
31. Challender D, et al. (2016) On scaling up pangolin conservation. Traffic Bulletin 28:
45. Wardle DA, et al. (2004) Ecological linkages between aboveground and belowground
19–21.
biota. Science 304:1629–1633.
32. Fennessy J, et al. (2016) Multi-locus analyses reveal four giraffe species instead of one.
46. Ceballos G, Ehrlich AH, Ehrlich PR (2015) The Annihilation of Nature: Human
Curr Biol 26:2543–2549. Extinction of Birds and Mammals (Johns Hopkins Univ Press, Baltimore).
33. Butchart S, Dunn E (2003) Using the IUCN Red List criteria to assess species with de- 47. Knoll AH (2015) Life on a Young Planet: The First Three Billion Years of Evolution on
clining populations. Conserv Biol 17:1200–1202. Earth (Princeton Univ Press, Princeton, NJ).
34. Gaston K, Blackburn T (2008) Pattern and Process in Macroecology (Blackwell Pub- 48. Barnosky AD, et al. (2014) Introducing the scientific consensus on maintaining
lishing, Hoboken, NJ). humanity’s life support systems in the 21st century: Information for policy makers.
35. Thomas JA (2016) ECOLOGY. Butterfly communities under threat. Science 353:216–218. The Anthropocene Review 1:78–109.
36. Régnier C, et al. (2015) Mass extinction in poorly known taxa. Proc Natl Acad Sci USA 49. Ceballos G, Ehrlich PR, Soberón J, Salazar I, Fay JP (2005) Global mammal conserva-
112:7761–7766. tion: What must we manage? Science 309:603–607.
37. Burkle LA, Marlin JC, Knight TM (2013) Plant-pollinator interactions over 120 years: 50. Brown IL, Ehrlich PR (1980) Population biology of the checkerspot butterfly, Euphy-
loss of species, co-occurrence, and function. Science 339:1611–1615. dryas chalcedona structure of the Jasper Ridge colony. Oecologia 47:239–251.
38. Ter Steege H, et al. (2015) Estimating the global conservation status of more than 51. Environmental Systems Research Institute (2011) Release 10. Documentation Manual
15,000 Amazonian tree species. Sci Adv 1:e1500936. (Environmental Systems Research Institute, Redlands, CA).
Downloaded by guest on September 20, 2021

E6096 | www.pnas.org/cgi/doi/10.1073/pnas.1704949114 Ceballos et al.