A new approach to map landscape variation in forest restoration success in tropical and temperate forest biomes

摘要

1. A high level of variation of biodiversity recovery within a landscape during forest restoration presents obstacles to ensure large‐scale, cost‐effective and long‐lasting ecological restoration. There is an urgent need to predict landscape variation in forest restoration success at a global scale. 2. We conducted a meta‐analysis comprising 135 study landscapes to predict and map landscape variation in forest restoration success in tropical and temperate forest biomes. Our analysis was based on the amount of forest cover within a landscape — a key driver of forest restoration success. We contrasted 17 generalized linear models measuring forest cover at different landscape sizes (with buffers varying from 5 to 200 km radii). We identified the most plausible model to predict and map landscape variation in forest restoration success. We then weighted landscape variation by the amount of potentially restorable areas (agriculture and pasture land areas) within the same landscape. Finally, we estimated restoration costs of implementing Bonn Challenge commitments in three specific temperate and tropical forest biome types in the United States, Brazil and Uganda. 3. Landscape variation decreased exponentially as the amount of forest cover increased in the landscape, with stronger effects within a 5 km radius. Thirty‐eight per cent of forest biomes have landscapes with more than 27% of forest cover and showed levels of landscape variation below 10%. Landscapes with less than 6% of forest cover showed levels of variation in forest restoration success above 50%. 4. At the biome level, Tropical and Subtropical Moist Broadleaf Forests had the lowest (12.6%), whereas Tropical and Subtropical Dry Broadleaf Forests had the highest (22.9%) average of weighted landscape variation in forest restoration success. Our approach can lead to a reduction in implementation costs for each Bonn Challenge commitment between US$ 973 Mi and 9.9 Bi. 5. Policy implications. Our approach ident