Secondary succession in swamp gallery forests along 65 fallow years after shifting cultivation

Understanding the resilience of ecosystems to different land uses is critical to elucidate the mechanisms un-derlying regeneration processes and plan more sustainable land-use systems. While these relationships are well understood in continuous tropical forests, they are still poorly comprehended in forest formations embedded in a savanna matrix, even though these forest systems are critical for conserving savanna biodiversity and water resources. We assessed the time and capacity of swamp gallery forests to recover after shifting cultivation, in the Cerrado (Brazilian savanna), investigating the role of environmental factors in community recovery. This shifting cultivation system uses slash and burn techniques and establishes drained fields of 0.3-1.2 ha which are culti-vated for 10 years, then left fallow for 10-65 years. To assess the resilience of these communities, we analyzed a chronological sequence of 28 secondary forests (1-65 years) and compared them with four old forests (reference ecosystems). We evaluated the trajectory and recovery time of community diversity, composition, and structure (height, canopy coverage, density, basal area, and biomass) as a function of fallow time, water table depth, and soil characteristics using multiple linear models. Forest fragments were considered recovered when their attri-butes reached the mean values of the reference forests. Most attributes are projected to reach the mean reference values of old-growth forests within nine decades, except for the proportion of non-pioneer species, canopy cover, and biomass. Water table depth influenced the trajectory and time required for recovery species richness and floristic composition. Soils with higher pH and higher nutrient availability positively influenced the recovery of non-pioneer species composition, while soils with lower pH and high organic carbon availability favored an increase in canopy height throughout succession. The disturbances generated by this shifting cultivation system do not prevent the recovery of at least 80 percent of the assessed community attributes in 65 years, suggesting that the resilience of secondary forests was not lost. Considering disturbance at similar scales, we recommend that these ecosystems be left fallow for at least 90 years to complete the recovery of ecological attributes.