Spatial dynamics of forest recovery after swidden cultivation in the Atlantic forest of southern Bahia, Brazil

A major challenge regarding secondary forest development relies on the understanding of processes affecting species composition and species spatial distributions. My research was designed to study how dispersal limitation affects species spatial distributions during secondary forest development in southern Bahia, a center of endemism and diversity for plants and animals in the Atlantic forest of Brazil. I used a chronosequence of secondary forest stands with similar history of past land use to quantify stand-level spatial variations of forest structure and composition and assess the mechanisms driving forest recovery. Fifteen pairs of mature and secondary forest stands following abandonment of swidden cultivation were randomly selected to represent three age classes: 11 yr (10--12 yr old), 24 yr (22--25 yr old), and 40 yr (37--43 yr old) with five replications each. All secondary forest stands selected were established next to a mature forest. At each pair of mature and secondary forest, a 120 meter long transect including 20 meters into the mature forest and 100 meters into the secondary forest was used to investigate the established vegetation, tree regeneration mechanisms, and the environmental variables that may affect forest recovery. The importance of regeneration mechanisms along the distance gradient was directly assessed by monitoring the soil seed bank, seed rain, and seedling bank input to forest recovery. The relative importance of biotic and abiotic filtering was estimated through changes in phylogenetic structure accompanying secondary forest development. In addition, measurements of changes in light and soil nutrient availability during stand development were correlated with life history traits of established vegetation to detect abiotic filtering.;The results showed changes in light and nutrient availability with secondary forest development. However, neither soil and light parameters differed along the distance gradient used in this study. Vegetation structure underwent predictable trends during stand development, with an increase in size and stratification but forest structure recovery was not affected by distance from mature forest. Trees common to mature forest stands successfully recruited in secondary forests, including several endemic and threatened tree species. Tree species richness and diversity recovered rapidly after land abandonment, reaching levels compared to those of mature forests in no more than 40 years. Plots close to mature forest edge showed a higher species richness and higher proportion of trees originated from seed rain while plots away from mature forest edge had lower species richness and higher proportion of trees originated from the soil seed bank. The abundance, richness, and diversity of seed rain increased with forest age. Seed traps close to mature forest edge received a higher number of large-sized seeds than traps away from edges but no differences in the number of small-sized seeds was found along the distance gradient. Tree species abundance and richness in the soil seed bank was similar along the distance gradient. The seedling bank of 11 yr old stands had the highest number of seedlings, the highest number of recruited seedlings, the highest growth rate, and the lowest number of dead seedlings per plot. Tree species richness in the seedling bank increased steadily with age in secondary forests.;The results of this study suggest that abiotic filtering promotes both phylogenetic clustering in young stands and phylogenetic overdispersion in mature stands. Life history traits of tree species such as N-fixation, dispersal mode, seed dormancy, ability to persist in the soil seed bank, and weight of 1000 seeds were strongly related to site characteristics in secondary forests, indicating that abiotic filtering plays an important role throughout forest development. Nonetheless, changes in the phylogenetic structure accompanying secondary forest development suggested an increasing importance of biotic relative to abiotic filtering after the stand initiation stage. Differences in species composition along the distance gradient resulted in within-stand variation in the strength of environmental filtering and competition, and ultimately, in community dynamics. Overall, this study highlights that dispersal limitation influences spatial and temporal aspects of community dynamics.