Wednesday, August 4, 2010 - 4:10 PM

SYMP 14-10: Gaps in our understanding of tree diversity in a tropical secondary forest

Joseph B. Yavitt, Cornell University, S. Joseph Wright, Smithsonian Tropical Research Institute, Dennis H. Knight, University of Wyoming, and Gerald E. Lang, West Virginia University.

Background/Question/Methods   I was an undergraduate student at UCSB in the 1970s, and Professor Connell's course in general ecology made at least two lasting impressions on me. The first was the way he carefully and critically presented the now well-recognized intermediate disturbance model of diversity. The second was the value of data obtained from long-term observations and experiments in permanent plots. Several years later I had an opportunity for fieldwork on Barro Colorado Island, Panama. There, in 1968, Knight and Lang had established a 1.5 ha permanent plot in 60 year old (at the time) secondary, lowland tropical forest, and they had periodically recorded abundance, distribution, size, recruitment and mortality of all plants greater then 2.5 cm diameter. I decided to extend the study by periodically mapping the dynamics of canopy gaps, i.e., a small-scale disturbance, and relate gaps to forest composition. The study is interesting because the whole forest was regenerating from agriculture, i.e., an extreme disturbance, during construction of the Panama Canal.

Results/Conclusions   There were no canopy gaps in 1968; however, gaps increased in frequency and size as the forest matured. Canopy gaps were spatially clustered, and thus there were gap-prone and gap-free parts of the plot. Occasional windstorms created the largest sized gaps. Small gaps (< 100 m2) were colonized quickly by in-growth of branches from surrounding trees and by advanced regeneration. Temporal trends in the forest include decrease density of individuals and minor change in diversity. Whether the various scales of disturbance, low recruitment rates, and slow growth rates in this forest characterize all maturing secondary forest is unclear.