One-third (34%) of all global forest lost between 2001 and 2024 is likely permanent – meaning trees in those areas are unlikely to grow back naturally, according to a new analysis by World Resources Institute (WRI) and Google DeepMind.
The impact is even more severe in tropical primary rainforests, where a staggering 61% of loss is tied to permanent land use change – a major setback for some of the planet’s most vital ecosystems for biodiversity and carbon storage.
Researchers also warn that while the remaining two-thirds of forest loss is typically linked to “temporary” disturbances like logging or wildfire – it can still have lasting consequences. Forests may take decades to recover. And even when they do, they don’t always return to full health.
The findings, made possible by an advanced AI model and satellite imagery and developed by Global Forest Watch, Land & Carbon Lab and Google DeepMind, offer the most detailed local, regional and global view to date of what’s driving forest loss – and what can be done about it.
“We’ve long known where forests are being lost. Now we better understand why,” said Michelle Sims, Research Associate at WRI. “This knowledge is essential to developing smarter actions at the regional, national and even local level – to protect remaining forests and restore degraded ones”.
The new dataset distinguishes drivers likely to cause permanent loss – such as expansion of agriculture, mining, infrastructure and settlements – which accounted for the 34% (177 million hectares) of global tree cover loss since 2001. Permanent agriculture alone made up 95% of that total. In tropical primary rainforests, drivers of permanent land use change drove an even greater share: 61% of loss (50.7 million hectares) – nearly the size of Thailand.
While the remaining two-thirds of forest loss stems from causes typically viewed as temporary – such as logging, wildfires, natural disturbances or shifting cultivation – recovery is not assured. Forest regeneration depends on how the land is managed afterward, the type of forest and the degree of environmental stress it faces.
“Just because trees grow back doesn’t mean forests return to their original state,” said Radost Stanimirova, Research Associate at WRI. “They might store less carbon, have fewer species or be more vulnerable to future damage. And climate change is making many natural events like fires and pest outbreaks more intense and frequent, which makes it harder for forests to recover.”
Drivers of tree cover loss vary around the globe
In tropical areas like Latin America and Southeast Asia, permanent agriculture is the dominant driver, responsible for 73% and 66% of loss, respectively. In temperate and boreal regions such as Russia and North America, wildfires – often triggered by lightning or human activity – and logging are the primary drivers. In Europe, 91% of tree cover loss is due to timber harvesting, much of it within managed forests where regrowth is planned.
Some drivers have an outsized impact in specific regions, even if they’re minor globally – driven by local land use, economic activity and governance. For instance, mining and energy drive less than 1% of global tree cover loss but caused 28% in Peru’s Madre de Dios region. In Colorado, climate-driven bark beetle outbreaks accounted for 27% of tree loss over two decades, even though natural disturbances like pests, storms, and floods make up just 1.4% of global loss. These are just two examples – many more exist around the world, each shaped by distinct local dynamics.
What this means for people, nature and climate
Permanent forest loss has serious and far-reaching consequences: reduced carbon storage, accelerated biodiversity loss and heightened risks to water and food security – all at a time when the planet is already facing a deepening climate and ecological crisis. Even temporary losses can be dangerous; depending on how forests recover, they may still lead to long-term ecosystem degradation and a decline in the critical services forests provide.
What needs to happen now
The new data marks a significant advance for forest policy, addressing a crucial gap in global efforts to halt deforestation by 2030. By pinpointing the underlying causes of forest loss in different places, it enables policymakers, companies and communities around the world to design more targeted, effective solutions to tackle deforestation.
Researchers at WRI outline a set of urgent, evidence-based recommendations tailored to the specific drivers of forest loss. Among other priorities, they call for stronger Indigenous and local land rights, designing policies that reflect local farming and land use realities, improving how forests are managed and monitored and ensuring infrastructure and agricultural expansion are guided by strong environmental planning.
For example, enforcing laws like the EU Deforestation Regulation and supporting smallholder farmers with sustainable alternatives are key to tackling agriculture-driven loss. Meanwhile, reducing wildfire risk demands ecosystem-specific fire management and early warning systems.