New Study Reveals How African Restoration Projects Spill Over into Surrounding Landscapes
Researchers from the University of Sheffield and University of Cambridge have conducted the most comprehensive study to date on how habitat restoration in sub-Saharan Africa impacts tree loss beyond project boundaries.
As global policy targets over 350 million hectares for restoration by 2030, the research, published in Nature Sustainability highlights a critical phenomenon known as leakage – where the impacts of a project spill over into the broader landscape.
The study, led by Xinran Miao, a PhD student from the University of Cambridge, assembled a geospatial dataset of nearly 130,000 restoration sites across 27 countries, covering approximately 6.8 million hectares. With the Sheffield team helping to develop the methods and models used to quantify the impacts of restoration both within and outside the projects themselves. By analysing decades of satellite data, the team assessed whether restoration was associated with ‘positive leakage’ (where tree loss declined in surrounding areas) or ‘negative leakage’ (where tree loss increased beyond project boundaries).
The findings reveal marked differences among restoration approaches. Natural regeneration and active restoration were associated with reductions in tree loss of up to 9.1 percentage points within project areas and 8.9 percentage points in the surrounding areas up to 10 kilometres away after ten years. This positive leakage may reflect projects that engage local communities, strengthen land tenure, or provide resources and livelihood opportunities such as fuelwood and jobs, thereby reducing the pressure on nearby natural forests.
In contrast, agroforestry – the integration of trees into croplands – was associated with negative leakage. On average, tree loss increased by 7.9 percentage points within these projects and 7.5 percentage points in surrounding landscapes. The researchers suggest this may happen when adding trees to active farms reduces agricultural production through competition for light or other resources, prompting farmers to clear nearby forest or fallow land to compensate for lost production.
Dr Oscar Morton, Leverhulme Early Career Research Fellow at the University of Sheffield, said: “These findings provide the first large-scale empirical evidence that different restoration strategies can generate heavily contrasting positive and negative effects on deforestation rates in the wider landscape.”
The study emphasises that restoration outcomes depend not only on what is planted, but also on where and how restoration is implemented. Results varied considerably among countries. Kenya and Tanzania showed consistently positive outcomes across restoration types, while Ethiopia and South Africa experienced instances of negative leakage, reflecting differences in socio-economic conditions, land-use pressures and institutional contexts.
“These results illustrate how factors such as land tenure security, incentive structures, and community participation can influence whether restoration strengthens or places additional pressure on surrounding landscapes,” said Dr Chris Bousfield from the University of Sheffield.
The findings have important implications for carbon markets and certification standards. Current frameworks often neglect off-site impacts, potentially underestimating the value of native forest restoration or overlooking risks associated with some agricultural and timber production interventions.