2018 ESA Annual Meeting (August 5 -- 10)

COS 69-1 - Seed to seedling transition shows distance-based mortality effects but no strong Janzen-Connell patterns for tree species at Barro Colorado Island (Panama)

Wednesday, August 8, 2018: 8:00 AM
338, New Orleans Ernest N. Morial Convention Center
Philippe Marchand1, Liza S. Comita2, S Joseph Wright3, Richard Condit4, Stephen P. Hubbell4 and Noelle G. Beckman5, (1)Institut de recherche sur les forêts, Université du Québec en Abitibi-Témiscamingue, Rouyn-Noranda, QC, Canada, (2)School of Forestry & Environmental Studies, Yale University, New Haven, CT, (3)Smithsonian Tropical Research Institute, Panama, Panama, (4)Smithsonian Tropical Research Institute, Panamá City, Panama, (5)Ecology Center / Biology Department, Utah State University, Logan, UT
Background/Question/Methods

Patterns of seed dispersal and seed mortality influence the spatial distribution of plant species and the local coexistence of competing species. Most seeds are dispersed in proximity to the parent tree, where mortality is also expected to be the highest, due to competition with siblings or the attraction of natural enemies. On the other hand, seedlings near the parent may benefit from root mutualisms or a more suitable habitat. The relative strength of these effects may produce different seedling establishment patterns from the seed source, including hump-shaped (Janzen-Connell) or declining (Hubbell) establishment patterns. This study investigates the shift in spatial patterns from the seed to seedling stage to explore the frequency of plant recruitment patterns proposed to promote local diversity through distance- and density-dependent mortality by specialized natural enemies.

In this study, we inferred the seed dispersal and seedling establishment curve for 12 tree species present in the Forest Dynamics Plot of Barro Colorado Island, Panama. We analyzed 13 years (2001-2013) of census data from seed traps, seedling quadrats, and adult trees located on the 50-ha plot. We fit hierarchical Bayesian models including parameters for seed production, the shape of the dispersal kernel, and the overdispersion of seed or seedling counts.

Results/Conclusions

For a majority of the modelled species, the seedling establishment curve decayed more slowly with distance than the seed dispersal kernel (i.e. Hubbell pattern), indicating a lower probability of establishment close to the parent tree. These results were not sensitive to the choice of dispersal kernel. Janzen-Connell patterns, where the peak of the establishment curve is at some distance from the parent tree, were identified for a few species, albeit with a large uncertainty in the shape of the kernel. Conversely, some species showed no evidence of differential establishment probabilities with distance.

Together, these results suggest that distance- or density-dependent mortality in the seed to seedling transition is not sufficient to produce Janzen-Connell density patterns for most species studied here. Few studies have focused on the transition of spatial patterns from the seed-to-seedling stage. We contrast this finding to other research at the same site showing that distance-dependent mortality does increase with time during the seedling stage. A close look at the timing of distance-dependent and density-dependent mortality in the tree life cycle may help elucidate the specific processes involved in maintaining spatial species diversity in a tropical forest.