SYMP 9-2 - Levels of ecological hierarchy clarify the determinism / stochasticity debate on tropical forest succession

Wednesday, August 14, 2019: 2:00 PM
Ballroom D, Kentucky International Convention Center
Sergio Estrada-Villegas1,2, Mario Bailón1, Jefferson Hall1, Stefan A. Schnitzer1,2, Benjamin L. Turner1, T. Trevor Caughlin3 and Michiel van Breugel4,5, (1)Smithsonian Tropical Research Institute, Balboa, Panama, (2)Department of Biological Sciences, Marquette University, Milwaukee, WI, (3)Department of Biological Sciences, Boise State University, Boise, ID, (4)Center for Tropical Forest Sciences, Smithsonian Tropical Research Institute, Balboa, Panama, (5)Yale-NUS College, Singapore, Singapore
Background/Question/Methods: A central debate in the study of ecological succession is the relative importance of determinism versus stochasticity as forest regenerate after severe anthropogenic disturbances. This debate, however, can be simplified by examining determinism and stochasticity across the levels of ecological hierarchy. Specifically, the relative importance of determinism and stochasticity may change predictably along a gradient of ecological hierarchy, with processes at higher hierarchical levels appearing to be more determinate and processes at lower hierarchical levels appearing to be more stochastic. At the ecosystem level, the mechanisms that control the increase in forest biomass as forests change over time may be determinate. By contrast, the change over time in species composition, a community property, may appear to be more stochastic. We surveyed 22 plots in a series of different-aged tropical dry forests in Panama for seven years to study how biomass and community composition of trees and lianas changed early in succession.

Results/Conclusions: We found that forest biomass increased predictably with forest age, indicating that deterministic factors control biomass accumulation in this tropical forest. At a lower hierarchical level of ecological organization, we found a much stronger role for stochasticity. However, the relative contributions of determinism versus stochasticity were not the same for trees and lianas: tree biomass showed a predictable increasing trajectory whereas liana biomass did not. Both tree and liana species composition showed no directionality, possibly due to the young age of the plots. To determine whether local variables increased predictability, we tested whether edaphic factors and initial conditions (soil nutrients, topography, forest cover at the time of abandonment, and basal area in the first census) affected the successional trajectories of trees and lianas. We found that edaphic factors made liana composition more predictable because a portion of the community seems associated with slopes and ridges. Initial conditions made the increase of tree biomass even more predictable because plots with lower initial basal area seem to experience less competition, allowing trees for accumulate biomass faster. Recognizing different levels of ecological organization may resolve why some features of early successional forests seem deterministic and others stochastic.