Tropical forests can recover more than 90 per cent of their original biodiversity within 30 years of agricultural abandonment, but the specific mix of species that defines an old-growth stand takes far longer, in some cases centuries, to return. That is according to a landmark study published this week in Nature, which analysed more than 8,500 species across 62 plots in Ecuador’s Chocó rainforest, one of the most biodiverse and heavily deforested regions on Earth.
Led by the Technical University of Darmstadt’s ‘Reassembly’ research group and involving 41 researchers from more than 30 universities and institutions across Germany and Ecuador, the study is the first to examine natural forest regeneration simultaneously across 16 taxonomic groups, spanning trees, birds, mammals, insects, soil bacteria, and leaf-litter arthropods, along a full regeneration gradient from active farmland to intact old-growth stands protected by the Ecuadorian conservation organisation Jocotoco.
The findings carry direct implications for global restoration targets. More than 60 per cent of tropical forests have already been lost or severely degraded, and the UN Decade on Ecosystem Restoration has placed secondary forest regeneration at the centre of international biodiversity commitments. Secondary forests, those regrowing after clearance, now account for roughly 70 per cent of remaining tropical forest area, yet their conservation potential has consistently been underestimated in policy frameworks laser-focused on primary forest protection.
Measured by abundance and species diversity, recovery was faster than researchers anticipated. Most taxonomic groups returned to near old-growth levels within three decades, with mobile fauna, including pollinators such as bees, frugivorous birds, and bats, showing the highest resistance to disturbance and the fastest return rates. Their early recolonisation of regenerating areas carries an outsized ecological effect, with seed dispersal and pollination setting the trajectory for wider recovery and creating a reinforcing cycle as returning vegetation attracts further animal communities.
Species composition, the specific identity of which organisms are present rather than merely their abundance, recovered far more slowly. Whilst abundance and diversity reached 90 per cent of old-growth benchmarks within 30 years, similarity in species composition reached only approximately 75 per cent of that threshold over the same period, meaning a regenerated forest and the stand it replaced may be structurally indistinguishable whilst harbouring fundamentally different ecological communities. Full compositional recovery can take several additional decades, and for some groups, considerably longer.
Professor Jörg Müller of the University of Würzburg’s Chair of Conservation Biology and Forest Ecology, whose team contributed acoustic surveys, camera traps, AI-assisted identification, and genetic species scanning to the research, said the new methods had opened a window into tropical recovery dynamics that had previously been impossible to study at this scale. “With these new methods, we can assess promptly the biodiversity of birds, mammals, frogs, and insects, as well as their recovery, even in the hyper-diverse tropical forest, at an entirely new level,” Müller said, confirming the group would continue developing the research over the next four years in partnership with Jocotoco across Ecuador.
Trees recovered on a substantially longer timeframe, constrained by slow generation times and limited dispersal capacity, whilst soil bacterial communities showed high resistance to disturbance but minimal evidence of return. Those communities remained measurably altered long after above-ground systems began resembling forest again, with leaf-litter arthropods and other less mobile organisms showing similarly persistent gaps in below-ground ecological function even where canopy recovery appeared structurally complete.
Prior land use proved a significant variable throughout the dataset. Forests regenerating on former cacao plantations recovered more quickly than those on former pasture, with plantation trees left on site providing shade and structural complexity that gave returning organisms a measurable head start over the competitive grass cover typical of abandoned pasture land. Recovery was also strongly sensitive to landscape context, with forests near intact habitat receiving a steadier influx of colonising species, accelerating regeneration, whilst isolated stands in more degraded settings showed slower and less complete recovery across almost all taxonomic groups.
Natural regeneration receives far less programmatic attention than active restoration, yet the study’s authors argue it can be both effective and cost-efficient, provided land tenure security allows forests to mature undisturbed. Many secondary forests are cleared again before they reach that point, short-circuiting their value as biodiversity reservoirs at precisely the stage when ecological returns begin to compound.
- The full study, “Biodiversity Resilience in a Tropical Rainforest” by Metz et al., published 8 April 2026, is available at doi.org/10.1038/s41586-026-10365-2.
