A suite of processes influence tropical forest vegetation dynamics during succession but the highly variable nature of tropical forest regeneration makes it difficult to generalize about the importance of chance vs. determinism in directing successional trajectories. Whereas these processes have been well documented in minimally disturbed forests, their impact on seed arrival and establishment in regenerating forests is not well resolved. We investigated factors influencing the seed-to-seedling transition (STS) using a large-scale tropical forest restoration experiment in southern Costa Rica. We asked: (1) how do restoration treatments, canopy cover, and life-history traits influence the STS; and (2) how do seed and establishment limitation influence local vegetation recovery across restoration treatments? We used a 3.5-year record of seed rain and seedling recruitment to examine limitations on the STS across natural regeneration and two active restoration treatments (applied nucleation and tree plantations) after a decade of recovery. For 43 focal tree species, we modeled the STS as the rate of seedlings recruiting per number of seeds arriving and determined how restoration treatments and species’ traits affected the STS. We used a randomization procedure to determine how local scale deterministic and stochastic processes influenced seed and establishment limitation across restoration treatments.
Results/Conclusions
Our results emphasize the importance of seed sources and species’ life-history traits in limiting tree seedling recruitment in degraded tropical landscapes. Consistent with previous studies we found that species with animal-dispersed seeds, and those with larger seeds (>5 mm), were more likely to make the STS. Furthermore, we found strong evidence that barriers to recruitment shift along succession, with the importance of seed limitation giving way to niche-based processes at later successional stages. This is illustrated by the fact that remnant reference forests were the least seed limited overall but had the lowest STS rates. Finally, while differential recruitment rates across the restoration treatments are likely mediated by differences in seed deposition rates, local scale dispersal patterns within treatment plots did not appear to limit recruitment. Overall, our findings demonstrate that catalyzing rapid development of canopy structure by planting trees not only decreases seed limitation by attracting dispersers and providing additional seed sources, but also alleviates establishment limitation by creating suitable niches for larger-seeded tree species. Moreover, after canopy cover is established, directly seeding larger-seeded species into plots could improve tropical wet forest restoration outcomes.