The 30-year forest: why the next big conservation idea is doing less

TL;DR

• A Nature study tracked 10,856 species and 23,590 bacterial sequences across recovering tropical forest in the Chocó lowlands of Ecuador, the most comprehensive biodiversity-recovery dataset ever assembled.
• After 30 years of being left alone, secondary forests regained >90% of animal abundance and biodiversity and ~75% of species composition compared with old-growth.
• The old consensus — that tropical recovery takes "a century or more" — was wrong by a factor of roughly three for most measures.
• Bees, bats, and frugivorous birds are the drivers, not the passengers, of recovery: they pollinate and disperse, the trees follow.
• The finding lands inside the UN Decade on Ecosystem Restoration, with 215 million hectares of tropical land globally identified as biophysically able to regenerate on its own — if simply left alone.
• The quieter story: 75% composition recovery means 25% of original species still aren't there at year 30. Some never come back without help.

A frog walks into a cacao plantation

In a thin strip of western Ecuador, sandwiched between the Pacific and the Andes, the Chocó lowland rainforest is one of the wettest places on the planet — eleven metres of rain in some years, more biodiversity per hectare than almost anywhere outside the Amazon, and a long history of being cleared for cacao and cattle, then quietly forgotten.

It is the forgetting that turned out to matter.

For three decades, ecologists working a chronosequence in the Chocó — paired plots of land at known stages of recovery, from active pasture to thirty-plus years post-abandonment to old-growth — quietly built up the most complete biodiversity-recovery dataset anyone has ever assembled. They counted trees, beetles, ants, butterflies, birds, bats, amphibians, reptiles, dung beetles, orchid bees, fungi. They sequenced soil. They listened. The full inventory ran to 16 taxonomic groups, three biological kingdoms, 10,856 named species or morphospecies, and 23,590 bacterial sequence variants — roughly an order of magnitude more taxa than any previous study of its kind.

The results, published in Nature by Timo Metz of the University of Marburg and a 30-strong author list spanning Ecuador, Germany, the United States, the Netherlands, Costa Rica and several other countries ¹, are unusual for ecology because they are unambiguously good news. After 30 years of being left alone — no replanting, no intervention, no management — the recovering forests had regained more than 90% of the animal abundance and biodiversity of neighbouring old-growth, and approximately 75% of the species composition.

"It's been a huge surprise for all of us," Metz told reporters via The Star wire pickup ². The previous consensus, drawn from decades of more limited studies, was that full tropical recovery took a century or more. The Chocó data say a useful 90% gets you there in 30 years.

That is the news. The interesting part is why.

The bees go first

Tropical forest recovery has, until now, been theorised mostly as a problem of trees. You plant the right species, you wait the right number of decades, you measure canopy closure. The new dataset reframes the whole thing.

What the Chocó chronosequence shows is that the mobile species — bees, bats, frugivorous birds — barely leave during agricultural use, and return almost immediately when pressure lifts. Their abundance had a median recovery time of 4.3 years in former cacao plantations and 25.5 years in former pasture ¹. These are the pollinators and seed-dispersers. They show up first, do the unglamorous logistical work of moving genetic material around, and the trees follow.

In other words: the bees are not waiting for the forest. The forest is waiting for the bees.

This inverts the long-running planning logic of tropical restoration, in which trees are the variable to manage and animals are the result. Metz and colleagues argue, with some force, that mobile animal communities are "drivers rather than passengers of tree recovery." Bats and frugivorous birds did the heavy lifting in moving large-seeded canopy species back into degraded land. The team showed mathematically that return rates of these animals contributed 1.0–2.5× more to overall recovery time than resistance (the fraction of species that simply hung on through the disturbance).

The implication, if it generalises, is that a useful restoration strategy looks less like a forestry programme and more like a wildlife-corridor programme. Protect the bats. Protect the bees. Protect anything that flies between an old-growth patch and a degraded one. The trees will get themselves back.

Where the numbers stop agreeing

Press coverage and NGO write-ups ³ have leaned hard on the cheerful headline — forests bounce back in 30 years, not 100. That part is real. But the dataset contains a less convenient finding too.

A median recovery is not a complete recovery. Across the 16 taxonomic groups, abundance recovered in anywhere between 0 and 258 years. Species composition — which species are present, not just how many — recovered in anywhere between 0 and 724 years, with bacteria taking even longer. The 90%/75% headline is an aggregate across taxa. Some groups are essentially back at year 10. Some are not back at year 250.

Specialist understorey insects, certain fungi, slow-dispersing amphibians, range-restricted lineages — these are the shadow species. They are also disproportionately the most endemic, the most threatened, the most ecologically unique. A 75% compositional match to old-growth means the regenerated forest looks like its old self from a distance, and very much does not from a distance of two centimetres.

There is also a methodological caveat the field will work through carefully. The Chocó is a single biogeographic region, with high connectivity between secondary patches and intact old-growth. Recovery in fragmented landscapes — Indonesian oil palm mosaics, Amazonian arc-of-deforestation cattle frontier, central African slash-and-burn matrices — may not match Chocó-grade rebound. The 215 Mha global potential figure that a related pantropical analysis produced ⁴ is biophysical capacity, not realised recovery.

The headline is correct. The headline is also not the whole story.

The economics this rewires

For most of the past decade, the dominant restoration framework — backed by the UN Decade on Ecosystem Restoration, the Bonn Challenge, REDD+ carbon finance, and most large NGO programmes — has assumed that active reforestation is necessary at scale. Plant trees. Buy land. Hire crews. Use drones and seed bombs and AI species selection. It is expensive. Estimates ran into the hundreds of dollars per hectare even before labour costs, before fencing, before the inevitable mortality of seedlings in degraded soil.

Active reforestation is still necessary in some contexts. But the Metz dataset is a serious thumb on the scale toward passive natural regeneration — letting land come back on its own — as the dominant strategy where it can work.

The financial geometry shifts in three places:

• Carbon markets. Forest carbon credits are priced partly on time-to-sequestration. If meaningful biomass and biodiversity recovery is 30 years rather than 100, secondary-forest credits become substantially more bankable. Existing methodologies — VCS VM0047, ART TREES — will need to be re-baselined.
• Restoration project finance. Cost per hectare for natural regeneration is dominated by land protection and opportunity cost, not labour and inputs. The cheapest tonne of CO₂ removed on earth may be the one you do nothing to.
• Conservation NGO portfolios. Organisations that spent the last decade building tree-planting operations will be asked, increasingly, whether the planting is doing useful work. Often the honest answer will be: not where land would have regenerated on its own anyway.

The early signal of this re-pricing is already showing up in indigenous-led conservation programmes, including a Wise Ancestors project in the Democratic Republic of Congo's Ingende Territory announced last week ⁵, which centres community land management and natural regeneration rather than plantation forestry.

The quieter story

What is genuinely new about the Metz paper is not the optimism. It is the methodology.

The Chocó dataset combines traditional field counts with environmental DNA (eDNA) soil sampling, bioacoustic monitoring of bird and insect calls, and AI-assisted species identification. It is what biodiversity science begins to look like when sensing catches up with field ecology. The team could plausibly not have produced this dataset a decade ago. The taxonomic breadth — 16 groups, three kingdoms — would have required hundreds of human-years of expert identification. eDNA and machine-classified bioacoustics compressed that into something a 30-author team could publish in a single paper.

This is the deeper signal. The forests recovering in 30 years was happening anyway. What has changed is that ecology can now see it happening, at a resolution that lets policy take it seriously.

What this means for you

The Chocó paper is a finding about land, but the audience that can act on it is narrower than the audience that will read about it.

• If you donate to conservation: ask whether your money is funding active reforestation in places where passive regeneration would work, and what mechanism the organisation is using to protect regenerating secondary forest from being re-cleared at year 10. Secondary forests are vulnerable precisely because they look unimpressive.
• If you work in land-use policy: the practical lever is turnaround time in forest management plans. Most tropical secondary forest globally is cleared again before it is 10 years old ¹. Extending protection windows to 30+ years is the single highest-leverage intervention the paper points to.
• If you are an investor in nature-based carbon: the 30-year recovery curve materially changes the discount geometry on long-dated forest credits. Re-baseline now, before the methodologies catch up.
• If you are a curious reader with no professional skin in this game: the useful thing to carry away is that the conservation framing you grew up with — the rainforest, once cut, is gone forever — was too pessimistic. Cut tropical forest, left alone, in the right places, comes most of the way back inside one human generation. That is not licence to cut more. It is licence to think differently about the cleared land that already exists.

Uncertainty ledger

• Generalisation beyond the Chocó. The dataset is a single biogeographic region with high old-growth connectivity. Heavily fragmented landscapes may recover much more slowly.
• The shadow species. Aggregate recovery percentages disguise large taxon-by-taxon variance. Endemic and slow-dispersing species may take centuries, or not return at all.
• Climate confounding. The 30-year window was measured under one climate regime. Whether recovery rates hold under accelerating warming, drought stress, and shifting rainfall is unknown.
• The protection question. Recovery is a biological capacity. Realising it depends on the political and economic question of whether secondary forest gets to stand for 30 years without being recleared. Most do not.

Bottom Line

The most important wildlife conservation finding in a decade is also the most boring. Tropical forests, left alone in the right places, recover most of their biodiversity inside thirty years — fast enough to matter to people alive now. The work the field still has to do is not figuring out how to make forests grow back. It is figuring out how to leave them alone for long enough.

Sources