Forest restoration requires strategies such as passive restoration to balance financial investments and ecological outcomes. However, the ecological outcomes of passive restoration are traditionally regarded as uncertain. We evaluated technical and legal strategies for balancing economic costs and ecological outcomes of passive versus active restoration in agricultural landscapes. We focused in the case of Brazil, where we assessed the factors driving the proportion of land allocated to passive and active restoration in 42 programs covering 698,398 hectares of farms in the Atlantic Forest, Atlantic Forest/cerrado ecotone and Amazon; the ecological outcomes of passive and active restoration in 2955 monitoring plots placed in six restoration programs; and the legal framework developed by some Brazilian states to balance the different restoration approaches and comply with legal commitments. Active restoration had the highest proportion of land allocated to it (78.4%), followed by passive (14.2%) and mixed restoration (7.4%). Passive restoration was higher in the Amazon, in silviculture, and when remaining forest cover was over 50 percent. Overall, both restoration approaches showed high levels of variation in the ecological outcomes; nevertheless, passively restored areas had a smaller percentage canopy cover, lower species density, and less shrubs and trees (dbh > 5 cm). The studied legal frameworks considered land abandonment for up to 4 years before deciding on a restoration approach, to favor the use of passive restoration. A better understanding of the biophysical and socioeconomic features of areas targeted for restoration is needed to take a better advantage of their natural regeneration potential.Abstract in Portuguese is available with online material.

The substitution of natural ecosystems with agriculture has led to the establishment of human-modified landscapes globally. In some tropical regions, this process is decades-old, allowing for the study of the effect of such modifications on the remaining biodiversity. However, unlike forest fragments inside regions with extensive primary coverage, the conservation value of ecosystems embedded within intensive farming, i.e., the anthropogenic matrices, has long been ignored, as have the effects of the landscape on such disturbed forest communities. Since the degradation process is predicted to cause the collapse of these fragmented forests, we can choose either to neglect them or to attempt the reversal of the degradation process for biodiversity conservation. Here we investigated the possible influence of landscape predictors on numerous plant species and on the relative proportions of different functional groups. Our results revealed that the richness found in human-modified landscapes had significantly more species than the protected reserves (+90%). The distribution of species suggested that any forest patch is likely to harbour a rare species. Generalised linear models and quantile regressions showed that forest cover and connected area influences the persistence of pioneer species and non-pioneer species of the canopy and zoochorics, with the latter also depending on slope. Rarefaction analysis revealed that natural remnants retain many species, even in sites with less than 20% forest cover. The presence of many zoochoric and non-pioneer canopy species may indicate a qualitative aspect to support conservation-restora tion efforts. These results indicate that the current strategy, which is limited to the preservation of biodiversity in public conservation reserves, should be reconsidered and should include smaller remnants of the natural ecosystem in a regional context and adopt large-scale restoration strategies to preserve the species pool.

No abstract