The Biodiversity Collapse: Why Species Are Vanishing Faster

Between 1970 and 2020, the average size of monitored wildlife populations fell by 73%, according to the WWF's Living Planet Report 2024. A separate 2019 assessment by the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) concluded that around 1 million animal and plant species are threatened with extinction. This article explains what those figures measure and what they do not, how fast species are really disappearing, which five pressures drive the loss, and which interventions have measurably worked.

Key takeaways

• The IPBES 2019 Global Assessment found that about 1 million species face extinction, and that the current extinction rate is "tens to hundreds of times" higher than the average of the past 10 million years.
• The Living Planet Index shows a 73% average decline in monitored vertebrate populations from 1970 to 2020. It tracks population size, not the number of species lost.
• The five direct drivers, ranked by IPBES, are changes in land and sea use, direct exploitation, climate change, pollution, and invasive alien species.
• About 75% of the land environment and 66% of the marine environment have been significantly altered by human activity.
• Conservation works where it is applied: one expert-elicitation study estimates it prevented 21 to 32 bird and 7 to 16 mammal extinctions between 1993 and 2020.

What biodiversity is and why it matters

Biodiversity is the variety of life at three nested levels. Genetic diversity is the variation inside a single species, which determines whether a population can adapt to a new disease or a warmer summer. Species diversity is the count and mix of distinct species in a place. Ecosystem diversity is the range of habitats, from coral reefs to peat bogs, that those species assemble into.

Think of an ecosystem as a power grid rather than a museum. A museum can lose a painting and the other paintings are unaffected. A grid that loses a generator, or a transmission line, may shift load onto other parts, and for a while nothing visibly changes. Then a heat wave arrives and the margin that redundancy was providing turns out to have been the whole safety system.

Ecologists call the useful outputs of that grid ecosystem services. They include the pollination of crops, the filtering of water by wetlands and forests, the formation and renewal of soil, and the nursery habitat for fish that people later catch. IPBES estimated that global crop output worth between US$235 billion and US$577 billion a year is at risk from pollinator loss. Those are the economic stakes of one service, and the figure carries a wide range for a reason: nobody can say exactly how much yield depends on each pollinator in each field.

Losing biodiversity is therefore not only an aesthetic or ethical problem, though it is both. It is also the gradual thinning of the redundancy that lets farms, fisheries, and watersheds absorb shocks. For a thought experiment on what happens when one group vanishes entirely, see our piece on what if all insects disappeared.

How fast are species actually disappearing?

The most quoted number in this debate is also the most misquoted. Older versions of this article, and many others, repeat that extinction runs "100 to 1,000 times" the natural rate. The more careful statement comes from the IPBES fact sheet: the current rate of species extinctions is tens to hundreds of times higher than the average over the last 10 million years, and it is accelerating.

Where does the "100 times" figure come from? A 2015 paper by Gerardo Ceballos and colleagues in Science Advances used deliberately conservative assumptions. The authors compared a background rate of 2 mammal extinctions per 10,000 species per 100 years with the modern record, and found that vertebrate species loss over the past century was up to 100 times the background rate. Their paper states that even under assumptions that would tend to minimize evidence of an incipient mass extinction, the losses are exceptionally rapid. You can read the abstract on the Princeton repository.

The table below collects the main headline indicators, with the body that produced each and the period it covers.

Indicator	Value	Source and period
Species threatened with extinction	About 1 million (animals and plants)	IPBES Global Assessment, 2019
Vertebrate species extinct since the 16th century	At least 680	IPBES, 2019
Amphibian species threatened	More than 40%	IPBES, 2019
Reef-forming coral species threatened	Almost 33%	IPBES, 2019
Marine mammal species threatened	More than 33%	IPBES, 2019
Average change in monitored wildlife populations	Down 73%	WWF Living Planet Report, 1970 to 2020
Land environment significantly altered	About 75%	IPBES, 2019
Marine environment significantly altered	About 66%	IPBES, 2019
Wetlands lost	More than 85% of 1700 extent, by 2000	IPBES, 2019

Source: IPBES media release on the Global Assessment (May 2019), ipbes.net; WWF Living Planet Report 2024.

One caution applies to the totals. Scientists have described roughly 2 million species, but the true number is unknown. The IPBES assessment worked with an estimate of about 8 million animal and plant species, and independent reviews note that methods give answers across a very wide range, from a few million to far more. That is why every global extinction rate is a bounded estimate and not a measurement.

What the Living Planet Index measures, and what it does not

The 73% decline is the most cited number in conservation and also the one most often misread. It does not mean that 73% of animals have died, and it does not mean that 73% of species are gone. The Living Planet Index (LPI) is an average of change across monitored populations. The 2024 edition drew on nearly 35,000 population trends covering 5,495 species of amphibians, birds, fish, mammals, and reptiles.

An analogy helps. A stock-market index summarizes the average movement of the companies in it, and a handful of enormous swings can pull the index a long way. The LPI works the same way: some populations have collapsed by 90% or more, some have held steady, and some have grown. The index reports the geometric mean of their changes, so heavy losses in a minority of populations drag the average down even when many others are stable.

The regional and habitat breakdowns in the 2024 report show where the average conceals the sharpest losses. Freshwater populations fell by 85%, terrestrial populations by 69%, and marine populations by 56%. By region, Latin America and the Caribbean showed a 95% decline, Africa 76%, and Asia and the Pacific 60%. The report also notes that some populations have stabilized or increased as a result of conservation action, which matters for the "what works" section below.

Two limits deserve mention. First, the index is built from vertebrates because those are the animals with long monitoring records, and it leans on well-studied regions. Second, it begins in 1970, which is not a pristine baseline: many populations were already depleted by then. The earlier figure of 69% that circulated widely, from the 2022 edition, covered 1970 to 2018, which is why the number has since moved to 73%.

Is this really the sixth mass extinction?

Paleontologists define a mass extinction by how much of life disappears over a geologically short span, and the previous five were driven by events such as flood-basalt volcanism and the asteroid impact that ended the age of the non-avian dinosaurs. For that last one, see our explainer on what killed the dinosaurs.

Whether the present qualifies is a live scientific argument, and honest reporting presents both sides. The 2015 Ceballos paper argues that the rate of loss already places us in the early stages. A 2022 review in Biological Reviews by Robert Cowie, Philippe Bouchet, and Benoît Fontaine, available on PubMed Central, extrapolated from mollusc data and argued that the Red List understates invertebrate losses by orders of magnitude. Yet the same authors say the current crisis has not yet reached the level of the previous five, and conclude that the sixth mass extinction has probably started. They also note that island species have suffered far higher rates of extinction than continental ones, and that the marine record does not yet show the same crisis as the non-marine one.

The practical reading is that "sixth mass extinction" describes a trajectory, not a completed event. Most of the damage so far is in abundance and range, meaning fewer individuals in fewer places. Extinction is the last step of that process, and it is irreversible, which is why conservation scientists focus on the earlier steps.

The five drivers of biodiversity loss

The IPBES assessment ranked five direct drivers by their relative global impact on nature over recent decades, in the following order. The ranking is for the period studied. It is not a forecast, and the weight of each driver differs by region and by type of ecosystem.

Rank	Driver	Representative evidence
1	Changes in land and sea use	About 75% of land and 66% of the ocean significantly altered; more than 85% of 1700 wetland area lost
2	Direct exploitation of organisms	33% of marine fish stocks harvested at unsustainable levels in 2015
3	Climate change	Roughly half of live coral cover lost since the 1870s (very wide uncertainty)
4	Pollution	Plastic pollution up tenfold since 1980
5	Invasive alien species	Major cause of extinction on islands

Source: IPBES media release on the Global Assessment, 2019.

1. Land and sea use change

Converting habitat is the biggest driver, because a species with no habitat cannot be rescued by any other measure. Most of the conversion is for agriculture, and the effects fall hardest in the tropics, where species are numerous and ranges are small. Our ranking of the countries with the most biodiversity shows how concentrated that richness is.

The Food and Agriculture Organization (FAO) estimated global deforestation at about 10 million hectares a year in 2015 to 2020, down from 16 million hectares a year in the 1990s. A UN glossary citing FAO states that an estimated 420 million hectares of forest have been converted to other land uses since 1990. That is a gross figure. FAO's net figure is smaller, about 178 million hectares, because forest has also expanded elsewhere through planting and natural regrowth. The two numbers answer different questions, and quoting the gross one as if it were net exaggerates the picture, while quoting only the net one hides the loss of old, species-rich forest that new plantations do not replace.

Fragmentation matters as much as raw area. A continuous forest can support wide-ranging predators and maintain the processes that keep populations connected. Cut it into patches and the same total area supports fewer species, because small patches cannot sustain viable populations of large or specialized animals. Our guides to deforestation and countries with the most forest cover go deeper on where the forest is going.

2. Direct exploitation

Hunting, fishing, logging, and the wildlife trade remove individuals faster than populations can replace them. IPBES reported that in 2015, 33% of marine fish stocks were harvested at unsustainable levels, 60% were fished at the maximum sustainable level, and 7% were underfished. In other words, only a small minority of stocks had room to grow.

Exploitation is the driver where regulation acts most directly. Quotas, closed seasons, and trade bans change the rate of harvest itself, so they can slow a decline faster than measures that must wait for habitat to regrow.

3. Climate change

Climate change ranks third in the IPBES assessment but is expected to grow in importance as warming continues. It acts through shifting temperature and rainfall, which push suitable habitat poleward or uphill faster than many species can follow, and through extreme events. For the physical science, see our explainer on climate change.

Coral reefs show the pattern clearly. IPBES reports that almost one-third of reef-forming coral species are threatened, and that roughly half of live coral cover has been lost since the 1870s, with a very wide margin of uncertainty. The IPCC's 2018 special report on 1.5 degrees Celsius projected that reefs would decline by 70% to 90% at 1.5 degrees of warming and that practically all of them, more than 99%, would be lost at 2 degrees.

Observed events are already severe. The Australian Institute of Marine Science (AIMS) reported in its 2024-25 condition summary that coral cover in the northern Great Barrier Reef fell from 39.8% to 30.0% in a year, in the southern region from 38.9% to 26.9%, and in the central region from 33.2% to 28.6%. AIMS described these as the largest annual declines in two of three regions in its 39-year record, driven mainly by heat from the 2024 mass bleaching, with cyclones and crown-of-thorns starfish contributing. The AIMS condition summary has the full regional data. For the biology of reefs themselves, read our explainer on coral reefs, and for the chemistry behind the second threat, see ocean acidification explained.

4. Pollution

Pollution reaches biodiversity through several routes. Fertilizer runoff drives algal blooms that strip the oxygen from coastal water. Pesticides affect non-target insects and the animals that eat them. Plastics, which IPBES says have increased tenfold since 1980, injure and entangle animals and add a chemical load to food webs. Our article on the plastic pollution crisis covers where that material comes from.

The insect evidence is the most striking. A 2017 study by Caspar Hallmann and colleagues in PLOS ONE analyzed 27 years of standardized trapping at 63 nature protection areas in Germany. It found a 76% decline in seasonal flying-insect biomass, and an 82% decline at mid-summer. The authors reported that weather, land use, and habitat characteristics could not explain the decline, and that it occurred across habitat types. The cause is still debated, and the study covers one country, but it is notable that the loss was found inside protected areas. The full paper is open access at PLOS ONE. The IPBES assessment gave a tentative estimate that about 10% of insect species are at risk, with a margin of plus or minus 10%, which reflects how poorly insects have been assessed.

5. Invasive alien species

When people move species across oceans, intentionally or not, the newcomers sometimes meet prey or competitors that evolved without them. Islands are the worst case, because island species often have no defenses against mammalian predators. Rats, cats, and other introduced animals have been linked to many island extinctions. Australia's government states in its Threatened Species Action Plan 2022 to 2032 that Australia has lost more mammal species to extinction than any other continent, and conservation scientists identify feral cats as a major driver. We discuss the protected-area side of that story in our ranking of countries with the most national parks.

The Amazon and the idea of a tipping point

A tipping point is a threshold beyond which a system shifts into a different state and does not easily return, the way a canoe tilts gently for a while and then suddenly flips. The Amazon rainforest has become the standard example. The forest recycles moisture through transpiration, so some of the rain that falls on it comes from the forest itself. Remove enough trees, and rainfall in the remaining forest can decline, which dries it out, makes fire more likely, and removes more trees.

A 2024 paper in Nature by Bernardo Flores and colleagues examined how three stressors interact: global warming, regional rainfall, and accumulated deforestation. Their projection is that by 2050, between 10% and 47% of the Amazon forest could be exposed to compounding stresses that might trigger irreversible local change. After crossing a threshold, the authors say, the forest may recover but stay in a degraded state dominated by opportunistic plants such as lianas and bamboo, or it may fail to recover and become an open-canopy, flammable ecosystem. The study stops well short of predicting collapse. It identifies where the risk of that outcome is highest and argues for a precautionary approach. Utrecht University summarizes the paper here.

The wide 10% to 47% range is itself the honest part of the finding. It tells you that scientists agree the risk is real and substantial, and disagree on exactly where the thresholds lie. That uncertainty argues for acting early, because a transition that is irreversible on human timescales cannot be undone by correcting the estimate later.

Why this matters even if you do not care about animals

There are three practical reasons that do not depend on caring about any individual species.

Food security. Crops rely on pollinators, soil organisms, and genetic diversity in wild relatives that breeders use to develop resistant varieties. The IPBES figure of up to US$577 billion in annual crop output at risk is one expression of that dependence. Our thought experiment on what happens if all bees went extinct shows which parts of the food system are most exposed and which are not.

Resilience to shocks. Diverse systems usually absorb stress better than uniform ones, just as a diversified investment portfolio is less exposed to one failing company. A single-species plantation or a heavily fished stock has less capacity to rebound from drought, disease, or a bad year.

Climate and water. Wetlands, forests, and soils store carbon and regulate flows of water. IPBES reports that wetland loss is currently about three times faster in percentage terms than forest loss, and that more than 85% of wetland present in 1700 was gone by 2000. Biodiversity and climate stability are linked problems, and our article on why coffee could disappear is one concrete example of a crop caught between them. Mining for clean-energy materials adds another pressure on habitat, which we cover in why rare earth metals are so critical.

What the evidence says about solutions

The most important finding in recent conservation science is that effort changes outcomes. A 2021 study in Conservation Letters by Friederike Bolam and colleagues used structured expert judgment to estimate how many extinctions conservation action has prevented. For the period 1993 to 2020, the estimate was 21 to 32 bird extinctions and 7 to 16 mammal extinctions prevented, against observed losses of 10 birds and 5 mammals. The authors calculate that extinction rates would have been roughly 2.9 to 4.2 times higher without conservation. They also caution that many of the saved species remain highly threatened. The abstract is indexed at DOAJ.

That result is an estimate based on expert elicitation, not a census, but it points in a useful direction: the species that recovered were helped by specific measures such as predator removal, captive breeding, and habitat protection. Here is how the 2022 global framework translates this into goals.

Kunming-Montreal target	What it commits countries to	By when
Target 2	At least 30% of degraded terrestrial, inland water, coastal, and marine ecosystems under effective restoration	2030
Target 3	At least 30% of land, inland waters, coasts, and oceans conserved and managed	2030
Target 18	Identify harmful subsidies by 2025 and reduce them by at least US$500 billion per year	2030

Source: Convention on Biological Diversity, Global Biodiversity Framework targets, Target 2 and Target 18.

Protection is the visible part of the plan, and the numbers show a gap. The Protected Planet Report 2024 found 17.6% of land and inland waters and 8.4% of the ocean protected, against the 30% goal. We look at how much of that protection is real in our guide to countries with the most national parks, including the finding that fewer than 5% of protected land has been assessed for management effectiveness.

The quieter and larger lever is the one the table lists third. Subsidies that encourage clearing, over-fishing, and fertilizer overuse work against every conservation dollar. Reforming them is less visible than creating a park, but it addresses the first and second drivers directly. Technology also helps with monitoring: satellite imagery and machine learning now track habitat change and poaching, a topic we cover in AI and biodiversity monitoring.

Common misconceptions

"Extinction only matters for charismatic animals." The loss that carries the most weight is often of organisms nobody photographs: pollinating insects, soil fungi, freshwater mussels. The services come from the whole community, and the least-studied groups are the ones where we know the least about what is being lost.

"The 100 to 1,000 times figure is a precise measurement." It is not. The estimates depend on assumptions about the background rate, on how many species exist, and on how many go extinct unrecorded. The range of tens to hundreds of times is the defensible summary, and it is large enough to be alarming even at the low end.

"A population decline means species are gone." A 73% average decline in monitored populations means abundance has fallen sharply, not that most species have vanished. Many declining species persist at low numbers, which is where recovery is still possible. The distinction is the argument for acting now rather than a reason for comfort.

"Conservation does not work." The Bolam study and the recoveries noted in the Living Planet Report both show measurable results. The problem is scale and funding, not feasibility.

What this means

The biodiversity crisis is better understood as the erosion of a system's redundancy than as a count of lost species. Each lost population, wetland, or forest patch removes a margin of safety that was invisible while it existed. The data are strongest on direction and weaker on exact magnitude: nearly all indicators point down, the largest driver is the conversion of habitat, and the uncertainty is wide enough to argue for caution rather than complacency.

The encouraging part is concrete. Where specific threats have been removed, populations have recovered, and the policy tools to reduce habitat conversion, over-harvest, and harmful subsidies are known and already written into an international framework. What remains uncertain is whether they will be applied at the scale and speed that the measured trends require.

Frequently asked questions

Q: How many species are going extinct each year? A: No one knows exactly, because the total number of species is uncertain and many extinctions go unrecorded. IPBES found that the current rate is tens to hundreds of times the average of the past 10 million years, and that around 1 million species are threatened. At least 680 vertebrate species have been documented as extinct since the 16th century.

Q: What does the 73% decline in the Living Planet Index mean? A: It means the average size of monitored vertebrate populations fell by 73% between 1970 and 2020, based on nearly 35,000 population trends for 5,495 species. It does not mean that 73% of animals or species are gone. Freshwater populations fell by 85%, and Latin America and the Caribbean by 95%.

Q: What is the biggest cause of biodiversity loss? A: IPBES ranks changes in land and sea use first, ahead of direct exploitation, climate change, pollution, and invasive alien species. About 75% of the land environment and 66% of the marine environment have been significantly altered. Agriculture is the dominant cause of habitat conversion, and it hits tropical regions hardest.

Q: Is the sixth mass extinction already happening? A: Scientists disagree about the label. A 2015 study found vertebrate losses up to 100 times the background rate. A 2022 review judged that the crisis has not yet reached the level of the previous five but that a sixth has probably started. Most agree the rate of loss is exceptional.

Q: Is the Amazon about to collapse? A: A 2024 Nature study found that by 2050, between 10% and 47% of the forest could be exposed to compounding stresses that might trigger irreversible local change. It does not predict collapse. The range reflects uncertainty about where thresholds lie, and the authors recommend a precautionary approach.

Q: Does conservation actually work? A: Yes, where it is applied. A 2021 expert-elicitation study estimated that conservation prevented 21 to 32 bird and 7 to 16 mammal extinctions between 1993 and 2020. The authors caution that many of those species remain highly threatened. Protection also lags its goal: 17.6% of land and 8.4% of the ocean were protected in 2024, against a 30% target for 2030.

Q: Can de-extinction solve the problem? A: Not at the scale needed. Cloning or genetic reconstruction of an extinct species would not restore the habitat or remove the threat that caused its loss, so it cannot stand in for prevention. It is best seen as a research field, not a substitute for prevention.

Sources

• IPBES, "Nature's dangerous decline 'unprecedented'" media release on the Global Assessment (May 2019): ipbes.net
• Ceballos et al. (2015), Science Advances, "Accelerated modern human-induced species losses: Entering the sixth mass extinction": collaborate.princeton.edu
• Cowie, Bouchet, and Fontaine (2022), Biological Reviews, "The Sixth Mass Extinction: fact, fiction or speculation?": pmc.ncbi.nlm.nih.gov
• Hallmann et al. (2017), PLOS ONE, flying insect biomass in protected areas: journals.plos.org
• Flores et al. (2024), Nature, "Critical transitions in the Amazon forest system," summarized by Utrecht University: uu.nl
• Bolam et al. (2021), Conservation Letters, "How many bird and mammal extinctions has recent conservation action prevented?": doaj.org
• Australian Institute of Marine Science, Great Barrier Reef condition summary 2024-25: aims.gov.au
• Convention on Biological Diversity, Kunming-Montreal Global Biodiversity Framework, Targets 2 and 18: cbd.int/gbf/targets/2, cbd.int/gbf/targets/18
• WWF and ZSL, Living Planet Report 2024: livingplanet.panda.org

Related Reading

• Deforestation Explained: Causes, Effects, and Real Solutions
• Why Coffee Could Disappear
• Climate Change Explained
• Why Rare Earth Metals Are So Critical
• Ocean Acidification Explained