Restoration of the Brazilian savanna after pine silviculture: Pine clearcutting is effective but not enough

Abstract

Planting tree monocultures in Brazilian savannas is a common practice for wood production. However, afforestation and fire suppression have caused significant ecological changes and biodiversity loss in the Cerrado region. Restoring these modified systems to their pre-plantation states remains a considerable challenge. In this study, we compared the recovery of savanna vegetation after afforestation in areas managed with three restoration strategies: (i) old pine plantation stands that are not managed anymore (control), (ii) passive restoration (natural regeneration after pine clearcutting), and (iii) native tree planting (native tree seedlings planted at high density after pine clearcutting). To compare the outcome of each strategy to native ecosystems that were never plantations, we sampled a cerrado sensu stricto (the desired restoration target), and a cerradão (the local alternative state when savannas are fire-suppressed). Fifteen years after applying the restoration approaches, we assessed the structure, richness, and composition of the plant communities. For the analyses, we separated the upper layer (woody native species, diameter at breast height (DBH) ≥ 5 cm) and ground layer (woody species with height ≥ 50 cm and DBH < 5 cm, and subshrubs and herbs), and distinguished savanna species from forest and generalist species. The results showed that both passive restoration and native tree planting restored the structure, richness, and composition of the upper layer, reaching the values of the cerrado sensu stricto. In contrast, in all treatments, the ground layer lacked the subshrubs and herbs typical of undisturbed savannas, with communities appearing similar to cerradão. Generalist species dominated all layers, and tree canopy cover in the treatments (mean values from 64.7% to 83.0%) ranged from 2.4 to 3 times the cover value recorded in the cerrado sensu stricto (mean of 28%). Pine clearcutting, independent of whether it was followed by native tree planting, improved the recovery of savanna woody vegetation. Therefore, we suggest that pine clearcutting, but not tree planting, is necessary to restore the woody community. Despite these apparently positive restoration outcomes for the woody plant community, it is important to note that after 15 years of passive restoration, communities are turning into low-diversity forests composed of generalist species and lacking savanna species in the ground-layer community. We therefore suggest that restoration of cerrado sensu stricto after pine clearcutting will require controlling woody encroachment, via prescribed fire or mechanical tree thinning, and planting of savanna subshrubs and herbs, rather than trees.