In the eastern Amazon, protecting standing forest outperformed restoration per conservation dollar
WHAT THE EVIDENCE SAYS
A Science study combined high-resolution satellite imagery with field data from 381 survey transects across 38 catchments and 499 rural properties, plus models for 160 bird species and 426 tree species. Using observed 2010–2020 land-use changes, the researchers compared counterfactual strategies for avoiding deforestation, preventing disturbance in standing forest, and restoring agricultural land. Avoiding disturbance produced the largest biodiversity benefits and was nearly as important as preventing deforestation for carbon; restoration remained important in heavily cleared properties but was less effective on its own across the full study area and was the most expensive strategy.
WHY IT MATTERS / OUR READ
The useful management signal is sequencing, not an argument against restoration. Where intact or degraded forest remains, preventing new loss and recurring disturbance can preserve more ecological value before managers pay to rebuild what has already been lost. Restoration becomes increasingly important where the landscape has already crossed that loss threshold.
CAVEATS / UNCERTAINTY
This was a large observational and counterfactual analysis, not a randomized intervention trial. Results come from two eastern Amazon frontier regions, so relative costs and returns should not be treated as universal; the study also evaluates modeled counterfactual outcomes rather than observing every strategy implemented side by side.
SOURCE
Science
View original evidence ↗
Stockholm Environment Institute · Research summary
View research summary ↗
Across more than three million hectares in Pará, counterfactual analysis found that preventing disturbance delivered the largest biodiversity gains, while restoration mattered most where forest loss was already extensive.
Read the Wild Lands assessment →