Abandoned pastures cannot spontaneously recover the attributes of old-growth savannas

Abstract

Active restoration strategies have been recommended to recover Neotropical savannas in abandoned lands, but no studies have investigated the trajectories and speeds of spontaneous recovery for these systems. Research into the dynamics of degraded savannas is urgently needed to guide restoration decision making. We analysed the dynamics of secondary savannas in the Brazilian Cerrado by sampling 29 abandoned pastures (time since abandonment ranging from 3 to 25 years) and applying the space-for-time substitution method. We modelled the temporal changes in plant community attributes and estimated the time (years) required for these attributes to match those of two reference ecosystems (three replicates each), old-growth savanna and a forest-type savanna, which had encroached following fire suppression (encroached savanna). We also analysed the plant community composition of the study sites. Our models showed that tree canopy cover, richness and density rapidly increased with time since pasture abandonment, easily surpassing the values of the old-growth savanna (28 years) and reaching the values of encroached savanna 49 years after abandonment. The cover and richness of the ground layer increased at a much slower pace. Since the species in this layer, including the exotic grasses, are shade intolerant, they will be eliminated by canopy closure over time. Up to 25 years after abandonment, secondary savannas continued to lack many (37%) old-growth savanna species, mostly from the ground layer (82% of grasses absent). This period was also not sufficient for the secondary savannas to become floristically similar to the encroached savannas, which are dominated by shade-tolerant tree species. Synthesis and applications. Despite the reported high natural regeneration of Neotropical savanna vegetation, abandoned pastures will not spontaneously return to an old-growth savanna state. Protected from fire and lacking the native ground layer, the end state of secondary savannas will be a low-diversity forest. If restoration goals include the recovery of old-growth savanna biodiversity and structure, interventions are required to prevent woody encroachment and reintroduce native grasses, forbs and shrubs. However, if the desirable endpoint is a low-diversity forest, passive restoration (non-intervention) and fire protection are appropriate.