Biodiversity has declined sharply since 1970, with monitored vertebrate populations falling about 73% on average and roughly one million species now threatened with extinction. Those losses are not uniform: they concentrate in fresh water, the tropics, and islands. The deeper story is about what those numbers measure, and why the next decade is a hinge point.
The machinery of loss
Habitat change is the leading driver, because it removes the physical template species depend on. When a forest is fragmented, edge effects penetrate inward: sunlight, wind, and invasive species alter conditions 100 meters or more from the boundary. Interior specialists then lose breeding territory even where total forest area looks stable on a map.
Overexploitation ranks second. Fishing and hunting can remove biomass faster than populations replace it, and large-bodied, slow-breeding species are hit first. Atlantic cod collapsed off Newfoundland in 1992 after decades of heavy harvest, and the stock has not recovered.
Invasive species and disease reorder ecosystems from within. Introduced predators on islands, and pathogens like chytrid fungus in amphibians, cause losses that are hard to reverse because the affected species have no evolved defenses.
Climate change is the escalating driver. Warming shifts suitable habitat poleward and upslope at rates many species cannot track. Coral reefs bleach when water stays roughly 1 degrees Celsius above summer maxima for weeks; repeated events leave little time for regrowth between them.
Pollution and land-use intensification compound these pressures. Nitrogen and phosphorus runoff fuels algal blooms that strip oxygen from coastal waters. Pesticides and light pollution then degrade the remaining habitat quality for insects and nocturnal species.
One plain-language point: biodiversity change is not one trend but many overlapping ones, and the overlap is where the damage multiplies.
The numbers behind the trend
The Living Planet Index tracks monitored wild vertebrate populations, not species counts. Its recent global estimate places average population decline at about 73% between 1970 and 2020. This is a mean of population trends, so it is not a share of species lost, and a small set of steep declines can pull it down.
The IUCN Red List offers a different lens: species-level extinction risk. Roughly 28% of assessed species are threatened, a figure that rises above 40% for amphibians and gymnosperms. The 2019 global assessment estimated about one million species threatened, a headline built from expert elicitation rather than direct census.
Extinction rate estimates are an order of magnitude softer than population trends. Background extinction runs at roughly 0.1 to 1 extinctions per million species-years; recent rates are often estimated at 100 to 1,000 times that, though the range reflects differing methods.
| Metric | What it counts | Approximate value | Key limitation |
|---|---|---|---|
| Living Planet Index | Mean change in monitored vertebrate populations | about 73% decline, 1970–2020 | Not species loss; skewed by steep declines |
| IUCN Red List | Species assessed as threatened | about 28% of assessed species | Taxonomically uneven coverage |
| Global assessment | Expert-estimated threatened species | about 1 million | Elicitation-based, wide uncertainty |
| Background extinction | Natural rate | 0.1–1 per million species-years | Hard to measure directly |
| Current extinction rate | Human-driven rate | 100–1,000 times background | Method-dependent estimates |
| Protected land | Terrestrial area protected | about 17% | Protection quality varies widely |
| Protected ocean | Marine area protected | about 8% | Enforcement often weak |
Boundaries, gaps, and who gets counted
Freshwater ecosystems lose biodiversity faster than land or sea. Wetlands and rivers hold a disproportionate share of species relative to their area, and water extraction, dams, and pollution hit them directly. The 83% average freshwater vertebrate decline often cited is the most extreme figure in the Living Planet Index.
Islands concentrate extinction risk. Endemic birds and mammals evolved without predators or competitors, so introduced rats, cats, and snakes devastate them. Most recorded bird extinctions since 1500 occurred on islands, which is why biosecurity on remote islands matters so much.
Timescale and geography shape what "change" means. A 50-year window captures industrial-era losses but misses earlier colonial-era deforestation. Tropical regions are now losing biodiversity fastest, while temperate regions show partial recovery from twentieth-century damage.
| Era | Dominant driver | Signature change |
|---|---|---|
| Pre-1500 | Land clearing, hunting | Local megafauna losses |
| 1500–1800 | Colonial trade, island settlement | Island bird and mammal extinctions |
| 1800–1950 | Industrial land use, overhunting | Temperate forest and large-mammal loss |
| 1950–2000 | Agricultural intensification, fishing | Freshwater and marine declines |
| 2000–2020 | Climate stress on top of land use | Coral bleaching, range shifts |
| Next decade | Climate plus habitat conversion | Compounding, faster change |
The counterintuitive finding is that some measures look better while others worsen. Local conservation and restoration have raised bird and mammal abundance in parts of Europe and North America, even as tropical declines accelerate. Aggregate global averages can therefore hide real progress and real collapse at once.
Misreadings worth correcting
Population decline does not equal species extinction. A 73% average drop in monitored populations says nothing directly about how many species were lost; most of those populations still exist.
Biodiversity is not just species richness. It includes genetic diversity within species and functional diversity, the range of ecological roles organisms perform. Both can erode while species counts hold steady.
Protected area percentage says little about enforcement. A park with no staff or funding may lose biodiversity as fast as unprotected land. Effective protection depends on management, funding, and local support.
More individuals of a few species is not recovery. Ecosystems dominated by generalists and invasives can look abundant while their resilience drops.
What practical action changes
Restore and connect habitat. Target areas large enough to sustain breeding populations and corridors that link them. Restoration takes years to decades; the first decade of a planting or rewetting project typically draws generalists before interior specialists arrive.
Manage what is caught and killed. Harvest limits, closed seasons, and gear changes can reverse overexploitation faster than any other driver, and the cod collapse shows the cost of delay.
Prevent and control introductions. Screen imports, and require ballast-water treatment for ships. Island biosecurity, including rodent-free protocols for supply vessels, is a well-established, commission-sized intervention.
Treat climate and biodiversity policy as one system. Protecting and restoring forests or wetlands stores carbon and supports species at the same time. Fund enforcement, not just designation. Budget for rangers, monitoring, and community programs where parks are declared. Use the right indicator. For extinction risk use the Red List; for population trends use the Living Planet Index; for ecosystem condition use national reporting. Red flag: a project that clears native vegetation should not be labeled biodiversity-positive on offset promises alone. Red flag: restoration that fails to specify reference conditions cannot be evaluated.
At small scales, the same logic holds. A garden that keeps native plants, leaves seasonal cover, and adds water supports more insects and birds than a lawn, and the effect compounds across neighboring properties.
What the evidence cannot settle
Global biodiversity indicators quantify different things. The Living Planet Index captures monitored populations, mostly vertebrates; the Red List captures assessed species; extinction-rate estimates derive from models. They agree on direction, not on exact magnitude, which is why headline percentages should not be reported as interchangeable.
Detection and effort bias the record. More survey effort and better reporting raise recorded decline and sometimes recorded recovery without any biological change. The mean-of-trends method also gives large declines outsized influence.
Future projections are scenarios, not forecasts. Estimated extinction risk under warming assumes emissions pathways yet to be realized. Keeping warming well below 2 degrees Celsius avoids the most severe coral and tropical losses, but local extinctions are already locked in.
What is certain is the mechanism: habitat, harvest, invasives, climate, and pollution interact, and their overlap drives the sharpest losses. What comes next depends on how aggressively those interactions are reduced, and how honestly the indicators are read.