Restoration can reduce tree loss nearby—or displace it beyond the project boundary
WHAT THE EVIDENCE SAYS
Researchers assembled 129,982 restoration polygons covering about 6.8 million hectares and used difference-in-differences designs to track tree loss inside projects and in surrounding buffers. Natural regeneration and active restoration reduced tree loss by as much as 9.1 percentage points inside sites and 8.9 points within 10-km buffers after 10 years. But outcomes varied sharply: some countries and approaches showed increased tree loss outside or even inside treated areas, including negative leakage associated with some active-restoration and agroforestry cases.
WHY IT MATTERS / OUR READ
The counter-signal is that leakage is not inherently negative. Restoration can relieve pressure in surrounding landscapes—but it can also move pressure somewhere else. Project success therefore needs a landscape boundary wide enough to detect both possibilities.
CAVEATS / UNCERTAINTY
The analysis uses remotely observed tree loss rather than a complete biodiversity or livelihood outcome, and causal identification depends on matching and difference-in-differences assumptions. Country-level estimates vary, restoration categories are broad, and results should not be converted into a universal ranking of approaches.
SOURCE
Nature Sustainability — primary study
View original evidence ↗
Across nearly 130,000 restoration polygons in sub-Saharan Africa, natural regeneration and active restoration often reduced surrounding tree loss, but country and treatment effects included clear cases of negative leakage.
Read the Wild Lands assessment →